Why does controlling thoughts and behavior feel effortful? Why do you need to pay attention? Why does this mental effort feel bad? Why do we feel drained after exerting mental effort and feel like we need to rest?

The most popular answer to these questions is known as the resource theory of willpower. The resource theory claims that there is some sort of resource that gets depleted every time you exert self-control, so that there is less available for exerting self-control in the future- a phenomenon known as "ego depletion." This aligns with (at least contemporary) common sense as well as numerous well-publicized studies and books that claim to see the effect everywhere.

The main test for ego depletion is called a sequential task paradigm. First you do one task that requires mental effort (say, solving puzzles), then you take a break, and then you do another that requires some other sort of mental effort (say, being awarded money and deciding how much to give to another test subject). The idea of ego depletion is that you will do worse at the second task than the first because you have used up your willpower juice.

The natural next question, then, is what is the resource being used up? For a while the most popular candidate was glucose, since taking glucose while doing an executive task generally was found to improve performance. But then another study found that just swishing glucose around in your mouth worked just as well. And then another study found that actually consuming glucose doesn't do much for performance, and furthermore that  "...exerting self-control is unlikely to reduce blood glucose levels... and performance on a second self-control task does not vary as a function of participants’ blood glucose levels." Nobody seems to have made a clear alternative proposal for what the resource might be.

Many individual studies have found evidence of ego depletion... but the most recent meta-analysis on the subject claims there is little evidence for its existence in general, and an attempt at replicating four ego depletion studies also failed to find an effect for any of them. This doesn't look great for the field.

It is hard to reconcile these studies with the common sense that everyone has at one point felt like their willpower was depleted by too many decisions. Furthermore, a friend of mine who majored in Brain and Cognitive Studies and has ADHD objected that these studies (presumably) only tell us about neurotypical people, and that neurodivergent people, specifically those with ADHD, do experience a genuine ego depletion effect. One possible way to reconcile these points is that different tasks impair people's willpower in varied (but not random) ways, and that these effects average out to zero when looking at a population with meta-analysis. But whether ego depletion is heterogeneous or nonexistent, one thing seems clear- the theory of willpower as a resource is insufficient to explain the data.


Here's a different question: why does physical effort feel effortful? You are hungry long before you are actually drained of nutrients, thirsty before you're dehydrated, exhausted from exercise before your muscles actually stop working. Yet the aversion that you feel is very real. The unifying trait of these aversive states is not that they signal a lack of resources, but rather that they motivate you to switch behaviors so that those resources will be freed up for other potential uses. It would be really bad if you exercised to the point where you actually couldn't run any longer and then had to dodge a predator later in the day. So the signal to stop needs to come long before the actual stopping point. 

Robert Kurzban says the thing managed by feelings of mental effort is not one that gets depleted over time, like gas or ATP. Rather, the managed resource is time itself. The mental effort associated with doing a task is a measure of the opportunity cost of continuing to spend mental processing time performing that task, relative to other options that are currently represented in your mind. The feeling of mental effort is a mechanism to convince you to spend your processing on some other task, in the same way that the feeling of physical effort is a mechanism to convince you to use your muscles for something else. Specifically, Kurzban proposes that there are regions of the frontal cortex that are "universal", in the sense that they can be used by any subsystem of the brain to achieve goals of importance to that subsystem, like playing piano or finding food or having sex. 

There are many decisions made by your brain that never feel effortful- for instance, your visual system has to constantly extract information about objects in the environment from raw visual data. Since there are multiple possible interpretations given a set of data, decisions are being made. But all this processing and decision-making doesn't feel effortful because the systems involved are single purpose. You can't use the neurons in your visual system for other tasks (at least on the time scales relevant to attention), so there's no point in a motivational state that tries to get you to use them differently.

Kurzban's idea fits well with various findings that giving more options makes people more anxious and less happy with their decisions. Even though you can often find a choice that fits your needs better, there will generally be only a small difference between whatever you picked and your second choice. In economic terms, you're more likely to be able to maximize utility, but the marginal utility is smaller. Unfortunately, your feeling of mental effort cares more about marginal utility, so you end up with this:

From John Campbell of "Pictures for Sad Children"
(Comic by Pictures for Sad Children)

The problem with technology is that its whole point is to make more options immediately available to you. Having a smartphone in your pocket changes your sense of options available to you, even when you're not consciously aware of it. There is always a subsystem in the back of your head, nagging your mind about giving it time and processing power to do what it wants. This is why I always make sure to both shut off my phone and leave it in my bag when rehearsing for a play. If your phone is in your pocket, even if you're not consciously thinking about it, it's still trying to get into your headspace.

In his book Getting Things Done, David Allen refers to these subsystems as "open loops." He describes ways to use what are essentially to-do lists to externalize these open loops. The most important function of the lists is that you trust yourself to check them regularly so these subsystems will be able to calm down and stop bothering your conscious mind. In this model, making one decision can actually make it easier to make more decisions in the future, as you no longer have the need to make the decision "hanging over you"- that is, constantly vying for the use of attentional systems. This definitely fits with how I feel on my (infrequent) productive days.

Because of the popularity of the resource depletion model, many people adopted the tactic of making as few decisions as possible to prevent "decision fatigue." Actually, it's not making fewer decisions that helps, it's being presented with fewer options. "Flow" states are ones in which your decision output is exactly matched with your option input such that everything "flows" straight through. If there's too much input, you get turbulence. Possible thoughts and actions keep presenting themselves, pushing to get processing time, and each time the thought is deferred, that's a bit of processing that is wasted since it doesn't lead anywhere. So attention is broken up and the sense of having to exert mental effort grows.

Now obviously, any account of opportunity cost will require a further unpacking of how the brain represents the costs of staying on the same task relative to switching. It's easy to see how effort doesn't seem to correlate to our conscious understanding of what's good for us. As Scott Alexander says, "[Kurzban's theory] strikes me as ridiculous – my brain is concerned that it has better things to do than study, but is perfectly happy with me playing Civilization IV: Fall From Heaven." We work nowadays under the assumption that the thing that our conscious mind thinks we should be doing is in fact the best thing that we could be doing. In a modern society, where you can provide for all of your needs by gaining comparative advantage through functional specialization, this is a defensible position. But we have no reason to believe this was true in the environment of evolutionary adaptedness. It takes a highly controlled, predictable environment for tasks with ambiguous, delayed rewards (such as studying) to be worth it.

As has been noted by others, the growing prevalence of ADHD may be simply a natural response to modern environments. More and more technologies are screaming to get inside our attention cycles, so it only makes sense that we would have difficulty paying attention to any single task. There are so many other tasks that are so appealing as to be barely distinguishable from other tasks. But this would predict that ADHD would be more prevalent in more urban areas, which is not the case according to the CDC:

Youth 4-17 who have ever been diagnosed with ADHD as of 2011, according to parent reports
I am admittedly very confused by this.


I learned recently that ADHD has a somewhat misleading name. In addition to having difficulty staying focused, people with ADHD will often experience hyperfocus, where they stay focused on a single tasks for extended periods of time while ignoring other important signals, such as hunger or needing to go to the bathroom. Taking this into account, ADHD isn't so much of a deficit disorder as it is a dysregulation disorder. Maybe for the DSM-VI they should change the name to "bi-focal disorder."

Since attention and mental effort seem to both revolve around the management of time (specifically processing time in the prefrontal cortex), we should suspect that effort and attention are also tied into our subjective sense of time. A friend with ADHD reports their sensation of time in different states as such:
Medicated: time is pretty consistent unless I hyperfocus (which can happen when I'm medicated but isn't as strong or as pleasant as when I'm unmedicated). 
Unmedicated and not hyperfocused: time is inconsistent and passes slowly in one sense because I experience every sound and sensation whether I want to or not but quickly in another sense because the smallest routine tasks take forever.

Unmedicated, hyperfocused: I'm incredibly into whatever thing I'm doing and time doesn't seem to pass slowly or quickly, I just seem to exist outside it.
This seems intriguing. One study had subjects tap to a metronome for a bit and then have them keep the same time without the metronome. Subjects with ADHD had more difficulty keeping consistent time: "While normal people lose their sense of the beat at 30 bpm, people with ADHD lose their sense of the beat at 40 bpm (see figure). This seems to suggest that people with ADHD experience the world on a faster time scale than other people."

On the other hand I imagine that there are plenty of trained musicians with ADHD who would do better than average on this task, so I'm not sure how seriously to take it. Another study shows a small effect size (d = 0.3) on "Piagetian time conservation tasks," which have something to do with noting that different things take the same amount of time, but the methods section isn't super clear.

Dopamine appears to be very important in regulating our internal clock rates, which is interesting since amphetamines and methylphenidates, two of the most common drug classes for treating ADHD, both function by making more dopamine available.

By testing the effects of various drugs on subjective time, researchers have pinned down the D2 dopamine receptor as key. Note that LSD, ketamine and salvia, all drugs known for screwing with the user's sense of time, are partial agonists of the D2 receptor. One hypothesis for why men have higher rates of ADHD than women is that men gain D2 receptors much more rapidly than women at the onset of puberty, which is when symptoms of ADHD typically start to appear. An imbalance between D2 receptors in the left and right hemispheres also predicts ADHD, and patients with initially high D2 receptor density respond better to methylphenidate treatment. D2 also mediates hyperactivity and response to amphetamines in mouse models.

The picture I have is that interactions between dopamine and the D2 receptor act as a "reward clock" tracking how much feedback you're getting over time. A mismatch between the two, driven by excess receptor production, makes rewards over different tasks less consistent. This results in difficulty focusing in some cases, and hyperfocus in others. Raising dopamine levels with medication brings the two components back in line with each other. In this sense, regulating attention is essentially regulating time.