These two questions arrive together from almost everyone who starts practising: how long until this does something, and will it stick if I stop?

The honest answers are unsatisfying, and giving them straight is more useful than an encouraging guess.

How long until it does something?

No published answer exists. Nothing in the ATT literature establishes a timeline from starting practice to a measurable change for a person practising at home.

What does exist is short and specific:

  • A single dose is enough to measure something. A randomized controlled comparison found increased resting alpha and beta oscillations in frontoparietal networks after one session[2].
  • Small numbers of doses have been used in performance studies. A sham-controlled study in healthy adults trained participants with two or four doses — two within one longer session, or four across two consecutive days — and found the ATT group significantly faster in correctly responding on a dichotic listening task[3].
  • The review across the field synthesised cognitive and neural effects across 14 attention tasks and three neural methodologies[4].

Note what that is and is not. Those are laboratory measurements taken close to the practice itself. “Something is measurable after one session” is a different claim from “you will notice a change after N weeks”, and nobody has published the second one.

Why the question is hard to answer even in principle

Two features of the protocol get in the way of the obvious answer.

There is no progress metric. The exercise produces no score, and it is designed to stay attentionally demanding rather than to become easier[1] — so “it got easier” is not available as a signal, by design. See does ATT get harder?.

Checking is discouraged. The pattern the technique targets is self-focused, threat-monitoring attention[1]. A daily habit of inspecting yourself for improvement runs that pattern. The manual’s measure of a session is whether you kept bringing attention back to the named sound[1] — a within-session criterion, not a weekly one.

So the exercise deliberately withholds the feedback the question is asking for.

Do you lose it if you stop?

Nobody knows. Follow-up measurements do exist. Haukaas and colleagues followed their student sample for six months[5], and a small 2026 mixed-anxiety trial assessed participants three and six months after its two-week intervention[6]. Those studies ask whether measured changes were still present later; they do not isolate the effect of stopping practice or establish how long any effect lasts without further sessions.

That narrower absence should be reported accurately. An answer in either direction — “you keep it” or “you lose it in a fortnight” — would be invention. The manual asks for regular scheduled practice[1], which tells you what the protocol assumes, not what happens when you stop.

What you can reasonably expect

Not a timeline, but a shape:

  • No emotional payoff per session. ATT is specified as not being a distraction, avoidance or symptom-management strategy, and not an emotion-management strategy[1]. A session that leaves you feeling much the same is behaving as designed.
  • Continued effort. It is meant to require practice and to stay demanding[1].
  • A dose the protocol names. Twice a day at about twelve minutes, with the manual’s own observation that in reality most people manage once[1].

The honest framing

The manual presents ATT as scheduled practice rather than a one-off intervention[1]. That describes the protocol; it does not establish a dose-response relationship or a timeline for benefit.

If you want a number, the only defensible one is the protocol’s own dose. Everything past that is not yet known — see does attention training actually work? for the full evidence position.

Questions and answers

How long before ATT starts working?
There is no published answer. Studies have measured effects after a single session in laboratory conditions, but nothing establishes a timeline for a person practising at home, and no study compares practice durations against each other.
Will I lose the benefit if I stop practising?
Nobody knows. Several ATT studies include follow-up measurements, but none isolates what happens because practice stops, so any confident answer — in either direction — goes beyond the evidence.
Should I expect to feel different after a session?
The protocol does not promise that. ATT is specified as not being a relaxation or emotion-management method, and its measure of a good session is whether you kept returning your attention.

References

  1. Wells, A. (2009). Metacognitive Therapy for Anxiety and Depression. Guilford Press. Chapter 4: Attention Training Techniques.
  2. Knowles, M. M., & Wells, A. (2018). Single dose of the attention training technique increases resting alpha and beta-oscillations in frontoparietal brain networks: A randomized controlled comparison. Frontiers in Psychology, 9, 1768. https://pmc.ncbi.nlm.nih.gov/articles/PMC6158576/
  3. Barth, V., Heitland, I., Krüger, T. H. C., Kahl, K. G., Sinke, C., & Winter, L. (2019). Shifting instead of drifting – Improving attentional performance by means of the attention training technique. Frontiers in Psychology, 10, 23. https://pmc.ncbi.nlm.nih.gov/articles/PMC6351493/
  4. Davison, C., Capobianco, L., Carter, K., & Wells, A. (2026). Cognitive and neuropsychological correlates of the attention training technique: A systematic review and evidence synthesis. Frontiers in Psychiatry, 17, 1766748. https://doi.org/10.3389/fpsyt.2026.1766748
  5. Haukaas, R. B., Gjerde, I. B., Varting, G., Hallan, H. E., & Solem, S. (2018). A randomized controlled trial comparing the attention training technique and mindful self-compassion for students with symptoms of depression and anxiety. Frontiers in Psychology, 9, 827. https://pmc.ncbi.nlm.nih.gov/articles/PMC5982936/
  6. Snuggerud, T. R., Nordahl, H., Vrabel, K. A., Hoffart, A., & Johnson, S. U. (2026). The feasibility and preliminary outcomes of a two-week attention training technique intervention for young adults with mixed anxiety disorders. Cognitive Behaviour Therapy, 55(3), 372–390. https://pubmed.ncbi.nlm.nih.gov/40424353/

Published by the makers of Heed, an app for practicing ATT-style sessions. iPhone · Android. Heed is a self-practice tool inspired by ATT research — not therapy, and not a medical device.