The Attention Training Technique needs something to attend to: several sounds, going at once, in different directions and at different distances. Wells’ manual specifies between six and nine sounds, depending on the level of demand required[1]. The arrangement matters as much as the count.

The specified layout

The manual’s structure is built around distance:

  • At least three actual competing sounds in the room itself[1]
  • A further two sounds nearby[1]
  • Two more sounds or locations in the far distance[1]

Those three bands are the point. An exercise in steering attention needs somewhere to steer it to, and near-versus-far gives the attention a dimension to move along that is separate from left-versus-right.

The silent ones

The detail people miss: not every location has to be making a sound. The manual describes some of them as potential sounds — a location you are directed to attend to, where the sound may not exist during the practice[1].

Being asked to attend to a direction where nothing is currently audible is not a fault in the setup. It is arguably the purest version of the task, since there is no sound doing the work of holding your attention for you.

How many is the right number?

The manual gives the range — six to nine — and ties it to the level of demand required[1], which makes the sound count the one openly adjustable dial in an otherwise fixed protocol. Published work has used the top of that range: a single-case study used three sounds in the room, three outside the room, and three in the distance — nine in total — with sessions of about twelve to fifteen minutes[2].

More sounds means more competition, which means more demand. Fewer means a task that is easier to hold together. Neither end has been tested against the other in a way that would let anyone call one better.

Real sounds or recorded sounds

In the original clinical setting, the therapist works with whatever the consulting room provides — the sounds that happen to be there, plus locations outside and further off. The manual also records that recorded versions of sounds have been used in the implementation of ATT[1].

That single sentence is what makes an audio or app delivery of ATT a version of the documented exercise rather than an invention. What a recording changes is control: the number of sounds, their placement and their character stop being a matter of what the room happens to offer.

What the sounds should be like

The protocol is far more specific about arrangement than about content. It does not prescribe particular sounds, and there is no published list of approved ones. What the structure implies is modest and worth stating plainly:

  • They have to be separable. The task is to hold one and switch to another, so two sounds that blur into one another give the listener nothing to move between.
  • They have to persist or recur. A sound that happens once cannot be returned to when the script names it again.
  • They should not tell a story. Anything that demands interpretation — speech you can follow, music you know — pulls attention into content, and the exercise is about the attentional process rather than the content of what you are attending to[1].

Changing them between sessions

This part is specified. Sounds and their arrangement are varied between sessions, which the manual describes as offsetting the effects of practice on task difficulty[1]. The exercise is meant to stay attentionally demanding rather than become substantially easier with practice[1], and a soundscape you have learned by heart works against that.

So there is no single correct set of ATT sounds — the protocol asks for the set to keep changing.

For what the voice does with these sounds once they are in place, see what the ATT script says. For the exercise in full, see the Attention Training Technique explained.

Questions and answers

How many sounds does ATT need?
Between six and nine, according to Wells' manual, depending on the level of demand required — at least three competing sounds in the room, two more nearby, and two in the far distance.
Do the sounds have to be real, or can they be recorded?
Both are documented. The manual describes using sounds in the consulting room, and also notes that recorded versions of sounds have been used to deliver the exercise.
What if one of the locations is silent?
That is intended. The manual calls some of them potential sounds — locations you are asked to attend to where nothing may be sounding at the time.

References

  1. Wells, A. (2009). Metacognitive Therapy for Anxiety and Depression. Guilford Press. Chapter 4: Attention Training Techniques.
  2. Carter, K. E. P., & Wells, A. (2018). Effects of the attention training technique on auditory hallucinations in schizo-affective disorder: A single case study. Case Reports in Psychiatry, 2018, 1537237. https://pmc.ncbi.nlm.nih.gov/articles/PMC6106726/

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.