What the research says about modern attention.

Prospective memory is your ability to remember to carry out an intention later. It is the cognitive function that allows the thought “I need to send that file after this meeting” to actually be executed after an interruption. This paper strikingly demonstrates that not every interruption is created the same.

Chiossi and colleagues randomized 60 participants to a 10-minute interruption involving TikTok, Twitter, YouTube, or no media, then tested whether they could correctly resume and execute previously formed intentions. The TikTok condition significantly impaired prospective-memory performance; Twitter, YouTube, and rest did not show the same effect. The authors attributed the difference to the combination of short videos and rapid context switching rather than video itself.

The raw result is difficult to dismiss. In the TikTok group, prospective-memory accuracy fell from about 80% before the interruption to 49% afterward. The endpoint was execution, or lack thereof: remembering what you meant to do after being pulled through a stream of unrelated content.

That is not a paper to be taken lightly.

This is one of the more convincing studies in the field because the investigators actually changed people's digital environment instead of relying on responses to questions asking how much participants used their phones.

In a randomized trial of 467 participants, mobile internet was blocked on participants' smartphones for two weeks while calls, texts, and internet access on other devices remained available. Objectively measured sustained attention improved by approximately d=0.24 in the initial intervention group. Mental health improved by about 0.57 SD and subjective well-being by about 0.46 SD. The delayed-intervention group subsequently showed improvements of similar magnitude when their own two-week block began. Overall, 90.7% improved on at least one of the three primary outcomes.

Compliance was imperfect, which we can understand just by looking around in public at all the heads down focused on phones. Only a minority maintained the block according to the study's strict compliance definition. Yet measurable benefits still appeared in the randomized analysis.

We don't read this as an argument for throwing your smartphone away. It is rather good causal evidence that constant access itself has a cognitive cost, and that at least some of that cost appears reversible within two weeks under the right conditions.

In two experiments, participants placed their smartphones either on the desk, nearby but out of sight, or in another room.

People performed worse on measures of working-memory capacity and fluid intelligence when their phones were more salient, even though they were not using the devices and did not report consciously thinking about them. The effect was largest among people who described themselves as more dependent on their phones.

Not every later experiment has reproduced every smartphone-presence effect, and one of Ward's own experiments did not find a sustained-attention difference. A 2023 experiment, however, again found poorer attentional performance from the mere presence of a smartphone.

The practical cost of putting the phone in another room when it's time to work is nearly zero. That makes even a modest effect interesting given the ease of the intervention.

For one week, 221 participants maximized phone interruptions by keeping notifications on and their phones nearby. During another week, they minimized interruptions by silencing notifications and keeping their phones away.

Participants reported more inattention and hyperactivity during the high-interruption week. Greater inattention was in turn associated with lower productivity and poorer psychological well-being.

This study does not say smartphone notifications cause ADHD. They do not. What it demonstrates is different but useful: an ordinary environmental manipulation can increase symptoms that look and feel like inattention and hyperactivity in people without a clinical ADHD diagnosis.

Turning off nonessential notifications is one of the cheapest cognitive interventions available.

A meta-analysis of 49 media-multitasking studies found an overall negative effect on cognitive outcomes. Across the cognitive studies, the pooled effect was approximately d=-0.71.

The penalty was larger when people had less control over the switching, when the competing tasks were unrelated, and when they were separated from one another. When people retained high control over the multitasking, the measured penalty was much smaller.

That distinction is particularly relevant now, as TikTok-style short-form feeds have become ubiquitous. Moving deliberately between planned pieces of work is not cognitively identical to being yanked unpredictably between Slack, email, a calendar alert, a text message, and a browser tab. Worse still is being tugged every 15 to 30 seconds toward an unrelated video.

There is not a problem with doing several things in one day. It is having the environment decide when your attention moves, rapidly and repeatedly, that is destructive to sustained focus.

Sophie Leroy gave a useful name to something almost everyone who does knowledge work has experienced: attention residue.

Across two experiments, people who switched away from one task continued thinking about it while performing the next. That residual cognitive activity impaired performance on the new task, especially when the previous task remained unfinished.

This helps explain why a two-minute interruption can cost much more than two minutes. The interruption ends. The prior cognitive state does not necessarily end with it.

A field study from Gloria Mark, et al showed how large that real-world disruption can become. They observed 24 information workers being interrupted. For interrupted work that was resumed on the same day, the average time it took before returning to the task was over 25 minutes.

“No More Side Quests” is partly a campaign line, but this is the underlying problem it describes. We've let side quests become the main quests, at least with respect to how much of the workday we allow them to consume.

This study examined computer-assisted protected time, where the calendar automatically reserved blocks for self-selected activities.

The authors first observed 27 existing users, then randomized 89 workers who had not previously used protected time. Workers given protected time reported improved performance, immersion, and job resources. Interestingly, their overall work activity decreased as their time was reorganized.

That is an important result. Better work does not necessarily mean filling more minutes with activity.

Protected time is one of the reasons Defragging Your Mind treats the calendar as an instrument for cognition rather than merely a record of meetings. Starting at the Bunny Slopes can change work lives much more than it appears on first glance.

A goal like “I am going to focus more tomorrow” has almost no structure. An implementation intention specifies the cue and action in advance: “At 8:30, I will put my phone away and work on the proposal for 30 minutes.”

Across 94 independent tests, these if-then plans improved goal attainment with an effect size of approximately d=0.65.

This is part of the reason we emphasize clear process goals, scheduled blocks, and predefined actions. The plan removes a decision from the moment when distraction is competing for the same mental resources.

This British Journal of Psychiatry editorial is provocative and should be read as a hypothesis-generating piece rather than established dogma.

The authors argue that excessive digital-media exposure may mimic or exacerbate ADHD symptoms and propose thinking about attention difficulties along a spectrum shaped by both genetic vulnerability and environmental exposure. They distinguish a traditionally neurodevelopmental presentation from a hypothetical environmentally influenced presentation, while explicitly acknowledging that the framework still requires empirical testing.

The timing of these ideas is quite relevant. National survey data estimate that 6.0% of U.S. adults, about 15.5 million people, currently report an ADHD diagnosis, and 55.9% of those adults received their diagnosis at age 18 or later. Those numbers do not establish overdiagnosis. They do establish that adult attention complaints and adult ADHD diagnosis now represent a very large clinical phenomenon.

ADHD is real, often impairing, and frequently persists from childhood into adulthood. At the same time, a person developing concentration problems in an environment of hours of daily screen time, constant notifications, poor sleep, anxiety, and repeated task switching deserves a broader assessment than a symptom checklist alone.

Cross-sectional brain-imaging studies always leave a difficult question: does heavy technology use change the brain, or do people with certain neural traits simply use technology differently?

It's a sophisticated chicken-or-egg problem.

Hu and colleagues tried to get closer to causality. They scanned 35 heavy and 21 light social-media users. The light users were scanned again after four weeks of experimentally increased social-media use.

After that month, the functional-connectivity patterns of the light users became more similar to those of the heavy users.

This was a small study and functional connectivity is not synonymous with cognitive impairment. We would not use it to claim that social media “damages the brain.”

It does support a simpler proposition: the functional organization of the brain is plastic enough to reflect repeated patterns of media use over surprisingly short periods of time. What we repeatedly do with our brains can show up as measurable changes in how neural networks function together.

A 2023 systematic review of 28 structural and functional MRI studies similarly found recurring differences involving reward, salience, attention, and default-mode networks in problematic social-network use, while emphasizing substantial methodological heterogeneity.