How Interruptions Destroy Focus: What Neuroscience Says About Deep Work, Productivity, and Learning

Once or twice per year, I like to read a productivity book to help hone my habits. I’m currently on Cal Newport’s “Deep Work,” which dives into the psychology of focus. It also offers some practical tips in the second half, which I haven’t gotten to yet. I want to take a moment to summarize and collect my thoughts on what it means to perform “deep work,” how it connects to “flow,” and how I can start better structuring my time.

The idea of “deep work” is a behavioral one: long stretches of distraction-free concentration on cognitively demanding tasks. Newport doesn’t go into how it feels to perform “deep work,” as he focuses on the idea of pushing your cognitive limits on something that matters.

Mihaly Csikszentmihalyi’s flow, on the other hand, is a psychological state. It involves total absorption, clear goals, immediate feedback, and a balance between challenge and skill. In flow, you lose track of time and self-consciousness and the activity feels intrinsically rewarding.

Neuroscience suggests these two ideas are indeed connected:

• Flow is linked to dopaminergic and noradrenergic systems that regulate motivation, mood, and the “drive” to stay engaged. 
• Deep, sustained attention recruits fronto-parietal and dorsal attention networks, which support maintaining focus, task rules, and goal-directed behavior.
• Flow tends to occur when people are pursuing learning progress at the edge of their abilities.

If you’ve lost track of time while programming, laying out a board, or studying a tricky engineering concept, you’re likely familiar with this state/feeling of “flow.” “Deep work” is the practice (and how you structure your time) of focusing your efforts, which ideally results in a flow state.

What your brain does when you’re deeply focused

Attention networks vs. the Default Mode Network

At a high level, your brain flips between two major modes:

1. Task-positive networks (fronto-parietal and dorsal attention networks) handle sustained attention, working memory, and goal-directed action. This requires active attention and using your working memory (e.g. writing a function while remembering why it’s needed in a codebase).
2. The Default Mode Network (DMN) is more active when you’re not focused on the outside world (e.g. self-reflection, autobiographical memory, future planning, daydreaming.

Deep work essentially means sustaining activation in the brain’s task-positive networks (primarily the fronto-parietal control network and the dorsal attention network) while preventing DMN-driven mind-wandering from constantly hijacking attention. But it’s important not to treat the Default Mode Network as a “villain.” While we’ll focus on the task-positive networks in this post, note that the DMN plays crucial roles in creative incubation, concept integration, and autobiographical processing. It becomes especially useful once you’ve saturated your task-positive networks with focused effort.

A useful way to frame this is not “task-positive good, DMN bad,” but “alternation between the two is essential.” Focused work recruits cognitive-control circuits to hold complex representations in mind. When you later relax your attention, the DMN becomes more active and supports associative processing, enabling your brain to reorganize information, form connections, and generate novel insights. This alternation aligns with decades of research on insight problem-solving: intense, effortful focus often precedes a period of diffuse attention, during which solutions or creative ideas “pop into mind.” In other words, deep work sets the stage, and the DMN often delivers the breakthrough.

When you’re “in the zone” debugging or writing, you’re biasing your brain toward the attention networks and keeping the DMN in check just enough that it doesn’t derail you with “what’s on Slack?” or “did I pay that bill?”.

Flow for learning

From a learning perspective, the interesting thing about flow is that it tends to emerge at a specific difficulty level:

• Tasks must be challenging but not overwhelming, sitting right at the edge of your current skill level.
• There is clear feedback (compiler errors, tests, plots, hardware behavior) that lets your brain constantly update its internal model.

Recent research suggests that flow isn’t driven by one big hit of reward; it’s more like playing a game of “hot and cold.” Your brain constantly evaluates whether your actions are bringing you closer to solving the problem or nudging you farther away. When those internal signals say you’re getting warmer (closing the gap between your current ability and what the task demands) it reinforces your focus and pulls you deeper into the work. That steady sense of “I’m moving in the right direction” is what helps sustain the flow state.

If you’ve worked on tough programming or engineering problems, you likely know what this feels like: it’s the “I’m so close” feeling that keeps you at your desk, forgetting to eat or go to bed.

Why deep work and flow feel so good

In addition to productivity, flow is strongly tied to subjective well-being:

• Flow states are often described as “autotelic” (i.e. the activity is its own reward). People in flow report enjoyment, reduced self-consciousness, and a distorted sense of time.
• Neuroscientific reviews connect intrinsic motivation and flow to dopaminergic circuits in the striatum and prefrontal cortex, linking progress and competence to positive affect.

What does this mean for us? Well, you know that feeling after a 2-hour coding session where you solved something hard and the world feels a bit sharper and lighter? That’s your reward system tagging the experience as valuable and worth repeating. But there are other benefits beyond just “feeling good:”

• Skill development: Deep work sessions create high-quality training data for your brain: long, coherent chunks of experience can be consolidated into more efficient internal models.
• Happiness and meaning: People who frequently experience flow report higher life satisfaction and engagement. For many engineers, the feeling of “I get to solve hard problems and it matters” is a core part of why we like our jobs.

How interruptions and multitasking break these systems

Sadly, modern knowledge work is basically designed to prevent these states. Let’s dive into some of these well-intentioned practices.

Multitasking wrecks learning and cognitive control

Psychologists have been studying task-switching and multitasking for decades. The picture is remarkably consistent:

• The American Psychological Association summarizes that switching between tasks incurs “switching costs.” These are measurable losses in speed and accuracy every time you change context, even when the tasks are simple. I liken this to context switching costs in a CPU.
• Research on media multitasking (using multiple streams of digital media at once) found that heavy media multitaskers show poorer performance on tasks involving working memory and cognitive control, including filtering irrelevant information.
• Reviews confirm that frequent media multitasking is associated with attention lapses, reduced memory, and altered cognitive control, though effect sizes vary.

From the brain’s perspective, constantly checking email, Slack, and social media trains the opposite habit of deep work. It rewards rapid context switching and shallow scanning instead of sustained, goal-directed thinking.

Interruptions in the workplace

When you move from lab tasks to real work, the picture doesn’t get better.

Erik Horvitz’s diary study of information workers at Microsoft found that people’s days are densely fragmented with interruptions and that resuming a task after an interruption often takes several minutes of re-orientation. This re-orientation time adds up and is associated with accumulated stress.

For developers:

• A mixed-methods study of 4,910 tasks from 17 professional developers plus a survey of 132 more found that task switching and interruptions impose cognitive load, degrade performance, and that self-interruptions (e.g., voluntarily switching tasks) can be even more disruptive than external ones.
• Recent work on “Breaking the Flow” in software engineering shows that interruptions during coding, comprehension, and review harm performance and that longer uninterrupted coding time is associated with higher perceived quality of the workday.

If you’ve ever alt-tabbed to “just check Slack” while debugging, you’ve likely felt this: it takes time to reload the full mental model of the system. The break in coherence makes mistakes more likely, and the work feels more frustrating and less satisfying.

Why this matters for learning

From a learning and neuro perspective, interruptions are toxic because they:

• Prevent stable engagement of attention networks, making it harder to build and maintain the rich internal models that real expertise requires.
• Increase mind-wandering and default-mode activity at the wrong time, which is correlated with worse comprehension and poorer academic performance when it happens during learning tasks.
• Train your brain’s reward system to seek small, frequent hits of novelty (notifications, updates, likes) instead of the slower, deeper reward of mastering a hard problem.

In other words: if your day is chopped into 5-minute slices by pings and meetings, your brain never gets to enter (or stay in) the states that best support learning, performance, and satisfaction.

Why this matters for engineers and students

For programmers, engineers, researchers, and students, the cost of shallow work is particularly high because your work:

• Has long dependency chains (a misunderstanding 3 layers down in your stack can invalidate hours of work).
• Requires large working-memory footprints (holding protocol specs, timing diagrams, call graphs, and constraints in your head at once).
• Often demands non-obvious abstractions and problem re-framing (the kind of creative leaps that emerge from sustained engagement, rather than from moving between contexts).

Deep work and flow give your brain the time and coherence needed to:

• Build chunked representations (e.g., recognizing patterns/bugs in code, schematics, and diagrams).
• Run longer chains of reasoning without forgetting critical assumptions.
• Experience the work as intrinsically satisfying, which is what keeps you coming back for more.

Things to try

I plan to create a full “how to” guide once I finish Newport’s “Deep Work.” Here are a few suggestions that I can come up with right now:

1. Schedule one deep-work block each day. Put it on your calendar. During that time, pick a cognitively demanding task (hard bug, paper draft, proof) and give it 60–90 minutes of uninterrupted time. Close Slack/Teams, email, social media, and silence your phone. Say “no” to meetings during this blocked time. Notice how long it takes to really “drop in” to the work and how you feel afterward.
2. Batch your shallow work. If possible, try to only check your email or Slack messages twice per day (e.g. once at noon and again at 4pm). Pick one day per week to handle all the “little tasks” (reports, blog posts, forum Q&A, email catch-up, etc.). On a personal note, this is the toughest for me, as I wander to these shallow communication tools while I’m waiting for something to compile or when I get tired toward the end of the day.
3. Reduce self-interruptions. When you feel the itch to do or check something mid-task (e.g. waiting for a compiler to finish) write it down on a sticky note. The idea is that you’re training your brain that it doesn’t need immediate novelty to feel safe.
4. Log it. If you’re a data-driven person, log start/end times of focused sessions for a week (I really like Toggl for this). Correlate that with how productive and satisfied you feel at the end of each day. You’re running your own tiny experiment on flow and deep work.

None of these require major life changes. If you had to pick one, I suggest going with the first option: schedule a 60-90 minute block on your calendar each day (if not once per day, then start with two per week). Do not allow anyone to put a meeting on your calendar during that time. Close out of all distractions and if possible, go to a place where people can’t interrupt you.

In addition to actually tackling work, view this time as “training,” just as you would for any sport or physical activity. You’re teaching your mind a new mode for focusing. During the beginning, you’ll likely slip up and check your phone, email, or social media. That’s OK. Simply recognize what’s happening, give yourself some grace, and return to the task at hand.

If you need to physically take a break or two, I recommend using something like the Pomodoro Technique with two or three 30-minute cycles (depending on the length of your block).

A couple of genuine deep-focus sessions per week can start to shift both your output and your relationship to your work. Neuroscience suggests your brain is built to reward you for it!

Recommended reading

Here are a few books that I’ve personally read and recommend if you want to dive more into the psychology of flow and productivity:

• Thinking, Fast and Slow by Daniel Kahneman
• Flow by Mihaly Csikszentmihalyi
• Deep Work by Cal Newport
• Algorithms to Live By by Brian Christian and Tom Griffiths
• Work Clean by Dan Charnas

How do you focus?

Please share your favorite productivity tips in the comments! I’m constantly looking for new ways to improve my work and flow, and I’d love to know what tricks or hacks you employ to help you focus.