Mindwandering

chapter one

ALWAYS “ON”

Much of neuroscience research up to the development of brain scanning was somewhat akin to phrenology, the Victorian practice of inferring people’s mental character by feeling the shape of their skulls. Of course, I’m overstating the case, but in studying the internal workings of the brain, the assumption for a long time has been that different areas within the brain would be dedicated to different tasks: one for language, another for memory; one for recognizing faces, another for feeling emotions. Over time, however, we came to realize that the operation and architecture of the brain are much more distributed over large networks than being modular and compartmentalized. Most, if not all, functions are accomplished via the activation and orchestration of multiarea networks. No single region, let alone individual neurons, accomplishes much without short- and long-distance cooperation. And in the context of mindwandering and the brain’s default network that mediates it, it is worth noting that different states of mind, such as meditation and sleep, as well as different psychiatric conditions all affect not only the information content in this massive network, but also the extent of connectivity between the cortical nodes of this network. The different areas constituting the network could be more strongly or more weakly connected in different states, more or less synchronized with each other, and influence each other to different degrees. Now we know that the brain is broadly dynamic and flexible in its operation and in its characteristics.

Still, we are far from a solid understanding of even the most basic neural functions. I learned this with a shock as a student in the laboratory of Professor Shimon Ullman, a pioneer in the development of computer vision. At the time, I was just finishing studying electrical engineering, out of an ill-conceived notion of fulfilling my father’s ambitions for me to be an engineer. I had quickly learned that I had absolutely no interest in chip design and that the only area of research in the field that captivated me was computer vision. The aim of this field was to mimic the way the human brain represents and recognizes images, and I discovered that at that time, thirty years ago, no one had a clear idea of how it is accomplished. I found that outrageous, and with the zeal of a young student with a whole lot yet to learn, I told Ullman so. He responded, as I recall, that I’d soon come to appreciate how complex the workings of the brain are. That I did. Sadly, it is still largely true that we have no hard knowledge of how the brain recognizes images, mainly some intriguing theories with preliminary support.

Fortunately, during my work in his lab, and then much more extensively in the cognitive psychology lab of another pioneer, Irv Biederman, a door was opened to a more productive, and exciting, new area of research that had just gotten under way, which I left to pursue. A major new way for studying the brain had recently been invented: fMRI (functional magnetic resonance imaging). The MRI machine itself, which uses magnetic fields and radio-frequency waves to image the anatomy of biological tissues, bones, and body organs, had been around for a few decades at that point, mostly in use in medical contexts. But the f, functional, MRI was the breakthrough neuroscientists had been thirsty for. By measuring blood flow, the functional part of fMRI allows us to infer where and when brain activity takes place. Maps of brain activity could be created by “sticking subjects in the magnet” and asking them to look at pictures, listen to sounds, count sheep—all sorts of tasks. We could look into the human brain during its normal, ongoing operation. Of course, that is with a few caveats, such as that what is measured is not exactly brain activation but rather a proxy, and that even interpretations of data can be subjective, but a revolution nevertheless. This was a moment of extraordinary adventure; we were roaming around inside the pathways of the mind like hikers in the woods at night with flashlights. And we soon stumbled upon the first truly substantial finding through neuroimaging.

The Discovery of the Brain’s Default Mode Network

Excited by the explosion of research, I got myself to Harvard Medical School, where Ken Kwong, Bruce Rosen, and collaborators were doing some of the most important work. My timing was fortuitous. A momentous discovery had recently been made: neuroimaging had paved the way to the discovery of the brain’s default mode and of the prevalence of mindwandering in everyday life.

What made the advent of fMRI so groundbreaking is that we no longer had to compromise on analogies to animals’ brains, we no longer had to make do with postmortem brains, we no longer had to infer the operation of the healthy brain from head and brain injuries (like that of the famous Phineas Gage or gunshot injuries in the Spanish Civil War), and we no longer had to limit ourselves to whatever can be recorded from patients during (or prior to) brain surgery. The result is beautifully colorful images that are taken as neural activation maps.

What are those colorful brain activations we see in fMRI studies? They are typically the result of a subtraction between what is evoked in the brain by two different experimental conditions. Imagine there is a study about emotional processing, specifically looking at what happens in the brain when we see happy faces compared with what happens in the brain when looking at sad faces. A participant (a “subject”) is asked to lie still on the sliding MRI bed—with a big cage (radio-frequency coil) around her head, loud high-frequency noises from the machine, in cold temperature—and attend what is projected to her on the screen. The fMRI signal is measured for each and every presentation trial. The brain activity elicited by all trials of one condition (all the happy faces) is averaged together and is subtracted from the average activation elicited by the trials of the other condition (all the sad faces). The resulting map shows areas where one condition activated more strongly (usually warm red-yellow colors) and less strongly (usually cold bluish colors) than the other condition. So, in our example, brain areas that show red are those where happy faces elicited stronger neuronal activity than sad faces, and blue spots correspond to regions where the sad faces activated more strongly than happy faces. And these maps are used to try to infer something new about the underlying neuronal mechanisms.

In between the experimental conditions (sad and happy faces in our example), there is a short rest period, usually a blank screen or a screen with a fixation dot at the center. This is used both for recovering the MRI signal for analysis purposes as well as to provide the participants with some rest in between experimental blocks of trials. And here comes the critical thing: While no one really believed that the brain is silent and inactive during those rest periods, the implicit assumption had been that it is much less active when our participants are resting and not required to perform any demanding task. The revolutionary discovery happened when, serendipitously looking at the activation maps during those rest periods, researchers started to notice that the brain is actually vigorously active when there is no specific task to perform, often more intensely than during the experimental conditions themselves, in a highly reliable manner, and in an extensive network, the DMN.

This accidental discovery of the DMN is most often attributed to Marcus Raichle and his coworkers, though the work was conducted by many labs.1 This network has since been dubbed the default network, this activity the default activity, and this state the brain’s default mode. Since its discovery, this default network has been readily found and replicated across many laboratories, paradigms, and MRI machines. By now it is accepted as a solid finding.

With all the excitement surrounding the early years of fMRI, it is clear now that what it measures and what is reported are not direct activations and not always consistent. There are many stages where distortion might creep in: from the moment we neuroscientists design an experiment to running it in machines that vary in sensitivities, with tens of parameters that could each change from one experiment to the next, to the analysis stage where there are numerous possible approaches with different strengths and weaknesses, to the limit of our interpretations. Indeed, the peak of the healthy skepticism regarding fMRI research matured in a recent study showing that when seventy different independent groups analyzed the exact same data set, they reported different results.2 While this is good to keep in mind as we are being exposed to more and more neuroimaging studies and subsequent claims, it is much less a concern in our context here. No one contests the existence and general behavior of the DMN. It is gigantic, it is omnipresent, and it is exceptionally replicable. We can proceed with our effort to understand the function(s) and characteristics of the brain’s default mode network.

The discovery of the DMN was sensational. Neural activity is highly energy consuming. Why would our brains waste so much metabolic energy when they were presumably doing nothing? When I arrived at Harvard as a postdoc, research had just commenced to determine what the function of the DMN might be. Using the intriguing method of thought sampling, combined with brain imaging, we learned that the more active the DMN is, the more one’s brain is engaged in mindwandering. It took the next couple of decades for the community to work out various important functions this seemingly spontaneous activity serves, with several different lines of research evolving.

As I grew as a neuroscientist, I came to develop two maxims about this fascinating research. The first is that evolution does not make mistakes. Everything we see in the brain has a reason and a function. Illusions, various “blindnesses,” cell suicide, false memories, and other findings that are puzzling and sometimes also amusing tend to make people believe they just caught the brain misbehaving, only to later realize these are various reflections of a grander strength. For the brain to be so flexible, adaptable, agile, and efficient, it has to pay some prices. (Indeed, when I am being asked why artificial intelligence algorithms do not behave like the human brains that they try to imitate, my answer is that AI is still more engineering than neuroscience. By making a computer perform a task with rigid boundaries of how to achieve goals, and with little accommodation for exceptions and improvisation, the artificial system lacks more implicit but immensely critical aspects of the human brain, such as flexibility and ingenuity.) So, in our context, once we realize that the brain is vigorously active when we are not busy with a specific goal, like when waiting in line, standing in the shower, or listening to something boring, knowing that this activity consumes significant energy should tell us that this activity must play some important role.

The second maxim, one that started in the mind of a young and naive postdoc but still serves me to this day, is that the brain always tells you, the inquisitive scientist, the truth. When things do not make sense, it is because you are not asking the right question or you are not asking the right question properly. The brain typically does not volunteer information, but the answers are there, waiting for us to arrive.

A relentless brain, always “on,” what does it do when we are not busy? The chapters ahead will tell the story of that often perplexing but always thrilling path of discovery and how findings that seemed quite disparate have been coming together. But before delving into this journey of unveiling the purpose of the DMN and of mindwandering, let us first give our thoughts a serious examination.

2

chapter two

CONNECTING WITH OUR THOUGHTS

We do not think about our thoughts often enough, but thoughts are the building blocks of our mental life, and of mindwandering. Thoughts are how we get from one idea to the next. They can be verbal, visual, and more; they can be progressing fast or slow; they can span many different semantic topics; they are based on stuff we know and have stored in memory; they can be of varying emotional valance; and they are often manifested as an internal dialogue between me and I. Thoughts are the interface and the translation of our inner world to our conscious mind, which can then be communicated to the outside world, or just remain with us.

The Source of Our Thoughts

When thoughts are aimed at a specific goal, they follow an agenda and a clear structure, not predictable, but nevertheless with a coherent progression, like when solving a problem. They accumulate and advance toward that goal. Planning is a good example. There is a chair at home you wanted to fix yourself, and you are thinking of doing it tomorrow morning. You think of the stuff you need to gather, like glue, a dead-blow hammer, a scrub plane, a wood chisel, a saw, and a sanding block. You “travel” along the web of concepts you have in memory, and you pick up the items that are relevant and necessary. You realize you need new protective gloves, so you think you will first go buy a new pair. You are thinking of where you will be doing the work exactly, the order of steps required for the repair, a complete simulation of what you are going to do to bring the chair back to function, how you are going to do this after everyone has left the house, and what your daughter’s reaction will be when she comes back to see her favorite chair fixed. That’s a train with a beginning and an end.

Sometimes we are more associative and more easily distracted, so when you get to the chisel in our list above, you diverge to thinking about Geppetto and Pinocchio and about a growing nose, and lying, and then you recall how your son told you he lied about having taken the dog out for a walk because he was lazy but too embarrassed to admit it. Then you think about how lucky you are to have this dog (and this son); you think about his playful attitude and how he lifts your spirits every day when you come back home. And you never get back to that fix-that-chair train.

We actually have one long train of thought during our waking hours. It changes topics, speed, style, orientation, content, and other characteristics, but it is continuous; there are no real pauses in thinking.

The source of our thoughts and what determines our next thought is a subject of developing and ongoing investigation. We all possess a sense of total agency over our thoughts, but this sense is unfounded. Conscious and subconscious thoughts mix, interact, exchange, and trigger processes in each other. We feel that we are privy to our thoughts and that if you asked us, we would know to tell you where our thoughts have originated and how one thought is connected to the one before and the one after. We believe that we are the owners and monitors of our thoughts, but this sense of agency is naive. You walk down the street thinking about an article you read last night, and suddenly you find yourself thinking about your high school teacher whom you have not seen for many years and who has no apparent connection to anything that was going through your mind a second ago. Our false sense of agency means that most of us, most of the time, cannot accept that the thought was triggered by a source unbeknownst to our conscious self. So we fabricate a link instead or just believe it popped up in our mind spontaneously.

But there is no such thing; thoughts do not just pop up. Each thought is connected to something, only sometimes this connection is beyond our conscious reach. That thoughts are connected does not mean that the thought process is always coherent, logically leading from one to the next. Interruptions to trains of thought could come from outside stimuli, like the sound of breaking glass or someone calling our name, or from an internal process, such as reminiscing on something emotional, and we can be aware of that interruption or not. Something on that street on which you were walking triggered the memory of your high school teacher, like somebody wearing distinct glasses like your teacher used to wear, but you could not trace back what that cue was. Perhaps it was because your eyes moved too quickly over that cue for the trigger to register in your conscious awareness or because you saw (or heard or smelled) something that you did not know in your mind was associated with your teacher. So now you are thinking about her without intending to and without understanding why, but your mind still follows that path.

Now let’s consider the path that our thoughts take even in a vacuum, with no interruptions whatsoever. Imagine the giant web that is your memory, with names, objects, places, concepts, and feelings all connected through associations. The thought process involves metaphorically walking on this web, from one node to the next, to transition from one concept node, or idea, to another. Every point in your path is connected to the previous one and to the next one, even if you cannot see it in every step. It is a web, so in every node you typically can go in multiple directions, and your mind chooses one. Let’s say you think that you need a vacation. When you stand on the “vacation” node of your web of possible thoughts, you could move on to the “money” branch and develop the financial consequences of going on vacation; you could go on the “fun” branch and start with happy mental simulations; or you could go in the direction of concrete planning of the right time and right destination for the vacation. With every step you take, your mind has to choose the next step out of several possibilities. Not consciously, and not with much deliberation, but it does. What determines the next step in your chain of thought is a little tug-of-war between different sources pulling you in different directions, and only one wins: your personality (frugal or not, open to new experience or not), state of mind, dispositions, the recent history of your thoughts (if you just paid the bills an hour ago you are more likely to go the “money” path, but if you just watched a commercial of a vacation on a beautiful island you will go on the “fun” path instead, owing to a phenomenon we call “priming”), or deep subconscious forces that draw you to the “I’ve got to get away from all this” path all compete for your selection.

While each node in our web of concepts and memories is connected to multiple other nodes at once, these connections are not of equal strength. Connections between neurons have “weights” that stand for the strength of their association. The strength between A and B determines how likely, how easily, and how quickly thought A will activate thought B, so when you see or think A, you will next think B. The strength of these weights may be determined by the quality of learning, such as how rehearsed this link is (how often a red light means “stop”), or could be temporary and determined dynamically by previous mental events that may keep a certain association primed.

Not knowing that activations of thoughts come from deterministic sources—be it priming by history, subconscious forces, or strength of association—can result, and does result, in many everyday confusions, not only our false belief in total agency over our thoughts and the illusion of free will that comes with it. Free association has been a major therapeutic tool since Freud and Jung and has proven potent in its capacity to unveil thoughts that are hidden from the individual’s conscious access. In the free-association method, the participant is presented with a word and encouraged to respond as quickly as possible with the first thing that comes to her mind, without censorship and without being judged. The idea is that under encouraging and nonintimidating conditions, inhibition is minimized and the things that then emerge from the freely associative responses are informative about the individual’s inner workings, deep desires, hidden fears, and surprising urges. But one needs to keep in mind the other sources that determine our next thought, as listed above, when trying to understand why I said or why I thought what I just did. If your therapist says “mother” and you respond with “blood,” she will be alarmed about your possible relationship with your mom. It could be a justified concern by your therapist, but another possible source for your response might be that you called your mom this morning to ask how to remove a bloodstain from your shirt and the semantic concept “blood” was thus primed, or warmed up, and therefore more readily provided as your quick response. We need to understand why thought A led to thought B before we can draw meaningful conclusions.

Observing Thoughts

My first formal adventure into my own inner world started by registering for a short course of mindfulness (mindfulness-based stress reduction, or MBSR, eight evenings and one final day of silence). There was a preliminary meeting to handle logistics such as paperwork and info on what to bring to the actual meetings. It was in Amherst, Massachusetts, where Jon Kabat-Zinn had fortunately distilled mindfulness meditation for the masses, and for this organizational meeting all groups were seated together in a giant circle on a basketball court. Just before standing up to leave, the instructors asked us to relax and close our eyes for one silent minute and then to share the experience with the others. It seemed like a benign, almost infantile, exercise. But this little minute opened a new world for me. The sudden pause, the radical change of my mental orientation inward, the long-forgotten sensation and attention to my own body struck me instantaneously. My life around that time was extremely hectic, being a young faculty in the competitive Harvard environment, with little kids at home, and more. When was the last time I had felt like that, I asked myself, and why had I not stopped for just one personal minute for so long? This is just like the clichéd question of when was the last time you looked up at the stars. Only this cosmos was inside of me, intense, personal, and waiting. I wanted to go deeper, and although it took me a few years to take that step, it paid off.

That I was encouraged to “observe” my thoughts initially sounded like complete nonsense. But I had made a decision before enrolling to suspend skepticism, to leave my scientist hat back home, and to come tabula rasa. So, I gave in and tried. We can observe ourselves in the mirror, notice a new wrinkle, focus on it for a little bit, and move on: observe, notice, examine, and let go. There is no real reason we would not be able to do the same thing inwardly on our thoughts. It is amazing how interesting, how accessible, and how intimate this experience can be, yet most of us do not venture in this direction our entire life.

We grew to see our thought process as impervious to personal examination. My first experience of silence, even as a (largely undisciplined) beginner, gave me the unequivocal realization that I had found a personal gold mine. Focusing on my thoughts very quickly started to seem like performing a sort of psychoanalysis on myself. Initially, you are occupied by mundane thoughts of what bothered you last: your trip over, things you left behind at home and at work, things you need and want to do after you are done, the smell in the room, or a distant sound. Then you start delving into older issues, memories, fears, and desires. You may find yourself smiling, or crying, with the sheer power of thinking inwardly. Strong emotions can be evoked simply by wading through our memories. And all this, I now know, can happen to you regardless of the mechanics of the meditation practice. The mere awareness that you are able and can benefit from looking into your mental being as a curious observer can then allow it to also take place during your everyday life, while making a salad or while jogging, and it does not require a special apparatus, clothing, or environment. The simple understanding that it is possible for me to look at my own thoughts has helped me gain a better notion of what bothers me, what makes me happy, why I say what I say, do what I do, feel what I feel, and behave like myself. That said, meditation can also expose thoughts and memories that you have been avoiding for a reason or that you do not yet have the skills to cope with and could use external help, so it is not a practice that is exclusively constructive.

I am obviously not the first one to stumble upon this fountain of observations about the self. Centuries of spiritual practices, psychological examinations, and even self-discoveries by many have preceded me. One that I would like to single out is Marion Milner (or Joanna Field, her pseudonym), who has gone long and far by keeping a meticulous and richly insightful diary when she decided to follow her experiences in a quest to find happiness. This eight-year journey is summarized exquisitely in her book A Life of One’s Own.1 Through her diary keeping, Milner developed a unique mastery of introspection. It is no wonder that she later became an established psychoanalyst.

We are used to being the subjects of our thoughts, being inside our train of thoughts, right at the center, almost as if our thoughts operate on us, with minimal control or insight of how and where they go. But this new practice of mine, and of many before me, means adding a vantage point of observing one’s own thoughts from the side. And this practice is really no art or something with which one needs tens of thousands of hours of experience before harvesting the benefits. It is merely an effort for a change of perspective. There are two possible perspectives: either you are inside your thoughts and experiencing them like a person sitting on a roller coaster, or you are observing them like someone who has not bought a ticket and is looking at the roller coaster from the ground. These two modes can then alternate, switching from immersive participation to outside observation, either automatically or at will. After a while it feels seamless to shift between them.

Integrating personal experience of observing thoughts with our progressive understanding of the mind (psychology) and the brain (neuroscience) allows a new and accessible grasp of who and why we are.

Thoughts and Mental Noise

In the world of engineering and signal processing, there is a measure termed signal-to-noise ratio (SNR). It quantifies how much a signal of interest is embedded within an otherwise noisy environment. Most realistic environments are incredibly noisy: radio reception in an environment filled with many other radio transmissions; visual images of the scene around us are crowded with clutter, obstructions, motions, varying illuminations, and more; and cocktail parties, where you have to fight distracting sounds and chatter to be able to understand what your friend is saying. A good system is one that amplifies the signal and suppresses the noise, maximizing SNR, so that you get what is interesting. The brain has to tackle the same issue, in considering both the external as well as the internal worlds.