Why Is Nobody Serious About Managing ADHD With Sound?

I like amphetamines as much as the next guy, but noise is well researched in neurotypicals and all kinds of attention spicy alike.

If you can’t manage attention, you probably know a bunch of possible explanations as of why you’re like that. The angle we’re interested in today is called

Low arousal theory

and it has been blessed with a very self explanatory name. The claim is that you are kind of clinically bored. There’s a lot free space between words you are hearing, number of objects in front of you is clearly too low and their movement is way too predictable.

‘Arousal’ has a long history as a research topic in psychology and I swear it’s useful to be aware of motivation for it. Somewhere between noticing that people are animals and that cocaine is absolutely amazing, the field got really into trying to figure out why people are different - thus bringing to life (reasonable approach to) the concept of personality. One of the first axis we got was extravert/introvert, and popular understanding of how that works was arousal: interacting with people is pretty information heavy, so it excites the subject. How much our subjects seeks it tells us how much excitement they are ‘lacking’, from which we get their default, background level of arousal. That allowed psychologists of the time to explain some individual differences between people.

Idea of ‘arousal’ was useful because, no matter how unreasonable it sounds, it generalized really well on many levels of complexity. As long as we were able to model something as a kind of living creature, we could define some sort of homeostasis it seeks and from that stimulate it less or more. Humans, dogs, neurons, rats. There were some bad turns, but as far as analogies go, it was a pretty good one.

One of the most fun places we applied it to was managing performance. You might have heard about ‘Flow’. If you have, you may not know it actually isn’t self help and coaching bullshit, but a totally serious theory with reasonable replication. If you haven’t, I’ll go over it and related/competing models for

Optimizing performance with stimulation

The Hungarian That Can’t Be Named made a bold claim, that sometimes people really get into doing something and they really enjoy that. The way to do it, according to him, is to do something you want to do, be at least kind of ok at it and keep it challenging. I can only assume your ADHD senses are tingling. Maybe some others too. Additionally, he wanted the task to have clear feedback and goals, but unfortunately that’s beyond the scope of this post.

Maybe it’s an isolated incident and there’s nothing there, so let’s grab another model. Yerkes-Dodson Law is one of my personal favourites in psychology. Just look at it (for the completely sane and normal problem of teaching mice picking white tunnel instead of black tunnel while receiving electric shocks)

I hope it’s just hard enough to keep you focused, but in case you have too much distraction, short explanation. For a simple task (learning to go into white passageway instead of black one when they are well lit) the harder mice were shocked, the faster they learned. At the time of authors writing the paper, that was the expected part¹: mice want to avoid stronger shocks more than they want to avoid weak shocks. If it's obvious and expected for you today, I'm sorry about your upbringing.

Interesting part, for the authors, was the harder task (passageways poorly lit). As it turns out, there’s optimal level of shock pain here. Mice get worse at learning when you shock them too much.

Nothing against flow (and I’m sure flow proponents are aware of what I’m about to type), but I prefer Yerkes-Dodson law because it gives us some very clear instructions on what to do

When task is too easy to focus

You can simply shock yourself harder. Voltage is a somewhat straight forward way of scaling how hard the task is. From flow perspective… what? Oh, you don’t want to shock yourself. Fair I guess.

Turns out, thanks to startlingly universal concept of 'arousal’, we can scale how hard the task is by combining core task and whatever additional cognitive load we want (plus some overhead per each ‘thing’, but lets not get into that'). Other than electric shocks, popular forms of load are:

• Physical activity

• Noise

• Flashing lights

• Smell/taste. Changing and incongruent, if you’re really feeling spicy

• Keeping things in (working) memory

So many options, which one to pick. As you might have figured out by now, I have a horse in this race. There's precisely one thing on that list that doesn’t affect your movement, doesn’t affect your primary sense, doesn’t require physical input and doesn’t habituate².

Here goes for all the ‘smart but lazy’ kids out there:

ADHD is when most tasks are too easy

Even though you’re somehow not even that good at them. Might seem weird, but that’s just my habit of looking for clickbait even in the middle of a piece. Correct term would be ‘understimulating’. Flow people would say ‘not challenging enough’, but that’s only because they’re misusing the term ‘arousal’ as seen in attached diagram

We got a bit away from ADHD into arousal research and sometimes pieces don’t fall in place when put together. Models can be good local approximations, but fall short from correct explanation. Does loading attention with noise help with ADHD?

There's pretty much no doubt. Maybe don’t overdo it with all senses, but other than that it very clearly works, even on top of drugs. I didn’t find any meta analysis, so you’ll have to take my word (or offer me a PhD track I guess) it’s not cherry-picked, but as far as I can tell the thing just works.

Which is probably as good of a time as any to say I have no answer to the question in title. I’m sure somebody is using noise in their therapeutic practice, but they don’t seem to be creating schools of psychotherapy around it. It’s really, really weird. That particular field is really into rotating fads and there is enough cognitive-behavioral psychotherapists who must enjoy arousal models as much as I do.

There is some problem with value capture from purely business perspective, but the same can be said about pretty much whole cognitive-behavioral therapy school and they don’t seem to losing members to starvation. Maybe it’s because people are self medicating - music is pretty cheap these days.

One other idea I have left is that maybe adjusting load level to real life settings is hard. That wouldn’t show in research: doing some basic pilotage to adjust task complexity to subject competence isn’t very hard. Plus most ADHD research is done on children and we have child developmental stages mapped really, really well. Lets hope it’s just a market inefficiency and get into

Dynamically adjusting noise to match task difficulty

I clearly run out of clickbait.

Adding flat amount of noise to your life is pretty easy. Here’s one page that plays ‘relaxing’(?) sounds to you, here’s basically the same thing with change in UI and here you can design your own ‘pure’ noise. I’m sure you’ve also heard about music, podcasts and playlists. So why have we all gathered here today?

All those noise generators are built with pretty static use cases in mind. Tinnitus, easier falling asleep, blocking out outside sounds or as a background noise for some predefined activity (cleaning, meditation, a third secret thing) and, honestly, are really simple designs. On/off button and volume slider per sound are really simple to setup and from what I can tell based on their development over years, target audience cares more about aesthetics (so they don’t immediately leave) or shareability, than about dynamic adjustment. Which is fine, I guess, but not what I care about.

Podcasts obviously don’t care at all and making music have the same general vibe over hours is hard enough, not to even mention I have no idea how would you make Spotify do anything other than switch tracks or change volume.

Which again, fine, both in general and for cognitive load, but under one assumption: your main activity has a constant level of challenge. Personally I will often want a bit more noise as I start compared to 5 minutes in as I tune in to the task. Furthermore, task difficulty tends to change - to give extremely relatable example, typical research paper will have much gentler introduction, than model discussion and trying to find issues in methodology is a separate kind of game.

‘You can switch sounds on/off on those pages you linked!’ I hear some say. You go and do that as you actually work on something. Removing sound all together is pretty easy (mute buttons are often on keyboards, headphones and speakers), but if you want to still have white noise to block your environment, but you want less chirping birds - tough luck, you’re opening site again. Cost of switching between tasks is beyond the scope of that, already kind of too long, text but you must have experience of answering a quick question as you’re trying to do something.

The part where I give you things to try out

Ideally we would want our sound to easily move on scale from less cognitively taxing to more - analogically to how you can move it from quiet to loud or switch between tracks back and forth. It would be really important that it does in the most accessible way possible, so you won’t lose focus from your main task.

I did some (personal) experiments with layering music, podcasts and repetitive noise on top of each other, since that was pretty easy to control with my setup and it was pretty good. Not mind blowing, it’s just changing cognitive load and I didn’t like white noise to begin with, but it was pretty good. You can do that, depending on your setup it should fine. But limitations were pretty clear and hitting wrong number of ‘tab’ after ‘alt’ is somewhat annoying.

So I made a thing. A free web app that lets you dynamically adjust noise load. You can control it only with built-in backward/forward track switching your setup provide you (any button with |« and »| ). You can even upload your own sounds and create sets, but they exist only locally in your browser and you can’t really share them. It even (mostly) works on mobile.

There’s also an open source app providing noise stimulation for attention issues. Runs on Windows right away, on Linux if you don’t mind running .py and anywhere else where you feel like running .py. You can switch between sets of sounds with hotkeys you can define without looking at it, you can add new sounds and sets of sounds, and even share them to some degree. I think it’s strictly better than web version, but I know people don’t download like it’s 1996 anymore.

So far I mostly layer pure BPM (120-180) on top of music and podcasts, but I’m aware most people want way less stimulation most of the time, so premade sound sets you can find assume you’ll be using them as only source of sound for the most part. But you can edit anything you want. I wish I could add more sounds but, you know, licensing.

Now go, do something. But don’t close that tab, so you can tell me how it went.

1. Teigen, Karl. (1994). Yerkes-Dodson: A Law for all Seasons. Theory & Psychology - THEOR PSYCHOL. 4. 525-547. 10.1177/0959354394044004.

2. Admittedly, you can get pretty far with the working memory thing. That said, I like to keep it open for adding internal singing, it’s pretty hype.