The cognitive shuffle is not just some random technique to facilitate sleep onset. There’s some serious thinking behind it. In fact, it is derived from a theory of sleep onset and insomnolence called “the somnolent information-processing theory”.

I first articulated this thinking in a paper simply published on SFU’s Summit archive. in 2013. In 2014 I published a shorter paper for the 2014 Cognitive Science Society’s Annual Conference’s workshop on “Computational Modeling of Cognition-Emotion Interactions: Relevance to Mechanisms of Affective Disorders and Therapeutic Action” in Québec.. My colleagues and I also presented the theory in a poster at the 2019 World Sleep Congress held in Vancouver, BC..

The theory and technique are outlined in more detail in this upcoming paper that has been accepted for publication:

Beaudoin, L.P. & Guloy, S. (in press). Towards a somnolent information-processing theory: Understanding the human sleep-onset control system from an integrative design-oriented perspective. In D. Kay (Ed.) The Cambridge Handbook of Sleep Theories and Models. Cambridge University Press.

It should appear in late 2025 or early 2026 (publishing of edited books takes time).

I expect to publish a preprint of our SIP article in PsyArXiv Preprints later this year. (I’m just waiting for clearance from the editor).

SIP theory paper abstract

Here’s the chapter’s abstract:

This chapter takes steps towards a Somnolent Information Processing (SIP) theory, developed from an integrative design-oriented perspective using the H-CogAff information processing architecture as background assumption. This provides an innovative lens through which to view sleep onset (SO) and is generative of new insights, strategies, and tools for sleep research and development in this area. Key to SIP is the recognition of the intricate interplay between affective, motivational, perceptual, and executive functions in governing the transition to sleep. SIP proposes six postulates to elucidate the regulation of SO and insomnolence. SIP highlights the role of chronobiological processes, mental perturbance (also known as “tertiary emotion”), computational alarms, and feedback loops in governing SO, while emphasizing the influence of learning mechanisms on somnolence. SIP postulates that while certain cognitive processes may delay SO (i.e., be “insomnolent”), others can facilitate it (“pro-somnolent”). It is hypothesized that, subject to sleep pressure and circadian timing and absent significant mental perturbance, being in a state that is similar in relevant respects to SO can facilitate SO.

The cognitive shuffle, serial diverse imagining (SDI) and serial diverse kinesthetic imagining (SDKI).

In the scientific literature, I tend to use the expression “serial diverse imagining” (SDI) rather than “cognitive shuffle” because it’s more precise. SDI is a type of cognitive shuffling. Cognitive shuffling is conjuring up and attending to diverse mental content items. Those types of content could simply be words, or they may be visual and/or kinesthetic items. If the content is visual or kinesthetic imagery, we call it SDI. If the imagery is specifically kinesthetic, we refer to it as serial diverse kinesthetic imagining (SDKI).

The cognitive shuffle elicited by mySleepButton is serial diverse imagining (SDI), including SDKI.

I think the first time in which I publicly used the expression SDKI is the SIP paper in the Cambridge University Press book referenced above. Well, it’s not quite public yet as that chapter hasn’t been published.

Examples of SDKI include imagining oneself

• walking in the woods
• playing the piano
• opening and closing drawers

There’s no end to the specific images one could conjure up when aiming to facilitate sleep onset.

What are some advantages of SDKI ( serial diverse kinesthetic imagining)

Theoretically, SDKI has the advantage of tying up more of the brain’s connectome, including major pathways in the connectome, compared to other cognitive bedtime techniques. The connectome is the system of neural pathways in a brain or nervous system, considered collectively.

If you imagine yourself opening a drawer with one hand and closing it with the other, there will be activity in the corresponding areas of your motor cortex and in the corpus callosum. The corpus callosum is the bundle of fibers that connects the two hemispheres (right and left halves) of your brain.

Why does this matter?

The more neural bandwidth is used by counter-insomnolent mentation (i.e., thinking that counteracts insomnia), the less bandwidth is available for insomnolent mentation (i.e., thinking that would keep you awake).

There are reports that as people fall asleep they sometimes feel themselves falling or moving around. SDKI might therefore be more pro-somnolent because it is more like natural sleep onset imagery. I.e., SDKI should be more counter-insomnolent and pro-somnolent than regular SDI. Having said that, we expect the optimal route to involve mixing up non-SDKI SDI with SDKI.

All this is pretty cool because our predictions distinguish the SIP (somnolent information processing) theory from rival theories. It’s often said that “cognitive arousal” (mental activity) is insomnolent. We disagree. We believe it all depends on the type of mental activity. Colin Espie and others often claim that the intention and effort to sleep is insomnolent. We disagree here too. It all depends on the form of the intention and the effort involved. All this is also cool because it distinguishes the cognitive shuffle from other techniques like monotonous imagery training, as discussed next.

Comparing the cognitive shuffle to other techniques

While the cognitive shuffle is based on a theory of sleep onset and insomnolence mentioned above (SIP) that taps into and extends prior research, it is also innovative. It calls for an empirical research programme. The theory raises a large number of new empirical questions. There have as yet only been preliminary empirical studies specifically on this theory.

It will require many experiments to tease out the diverse implications of this new theory. We see an acute need for studies comparing different forms of the cognitive shuffle with other bedtime cognitive techniques such as the following

• Monotonous imagery training (Morin & Azrin, 1987). Here participants are asked to mentally focus on a particular image for a couple of minutes. The “training” aspect is due to subjects practicing this imagery during the day. I have claimed that this is not particularly counter-insomnolent because it’s hard for participants to keep their mind focused on one boring image for two minutes when they have competing, insomnolent, thoughts such as interpersonal conflicts or financial problems. The cognitive shuffle is much more engaging. Other sleep theories would say that being engaging is a bad idea, but we beg to differ. Also, SIP theory implies that SDI is pro-somnolent because it (admittedly vaguely) resembles natural sleep onset mentation (imagery rich mind-wandering).
• Cognitive Refocusing Treatment for Insomnia is cognitive treatment proposed by Les Gellis. It encourages participants to shift their thinking away from insomnolent mentation (worries) towards neutral or pleasant thoughts. For example, one might imagine oneself walking around the house, or decorating a room. This technique has received only weak support. We think that is because like monotonous imagery training, it’s not particularly counter-insomnolent. If you have significant worries it’s hard to keep your mind focused on an alternative train of thought. The cognitive shuffle in contrast (particularly with the mySleepButton app) makes it easier to generate alternative images because the app speaks content to you. Also cognitive refocusing content is relatively coherent, whereas normal sleep onset tends to be more dreamlike involving mind wandering. SDI again would be more counter-insomnolent and more pro-somnolent.

So while the competing techniques have merit, our theory predicts that the cognitive shuffle will on average over a population of subjects be more pro-somnolent and more counter-insomnolent. This is a statistical prediction. There will of course be plenty of individual differences and other factors at play. Sometimes some participants will do better with the other cognitive techniques. Sometimes none of the cognitive techniques will work (e.g., if sleep pressure is very low, or subjects are very alarmed).

• Articulatory suppression is a more interesting cognitive technique that has received better empirical support than monotonous imagery training and cognitive refocusing (as systematically reviewed in a paper of ours.) This technique is very simple. The participant merely needs to continuously repeat a phoneme, such as “b”. This ties up the phonological loop of one’s working memory. This makes less cognitive capacity available to one’s worries. In other words, this is likely to be counter-insomnolent. This is likely more counter-insomnolent than monotonous imagery training and cognitive refocusing. However, it’s not likely to be pro-somnolent, because it does not share features with normal sleep onset (not much variety, does not involve visual imagery) — we don’t naturally fall asleep that way. Also, we expect it to be, on average, less counter-insomnolent than SDI because (for example) it does not tie up visual working memory (i.e., the visual-spatial scratch pad of working memory).

In sum, SDI and SDKI share some of the helpful features of the latter three techniques (e.g., tying up the visual-spatial scratchpad and the phonological loop) but goes beyond them (e.g., attempting to mimic sleep onset mentation).

Where’s the data?

As I have often said, as of now there’s very little data on the cognitive shuffle or the other cognitive techniques for sleep onset for that matter. It’s weird that something so important has received so little attention.

Even counting sheep, which is often said by sleep researchers to be useless, has received almost zero empirical attention. Yet there are reasons to believe that counting sheep may be mildly counter-insomnolent.

My colleagues and I have conducted three empirical tests of the cognitive shuffle two of which we have presented at conferences:

1. Digdon et al. (2016) Serial Diverse Imagining Task: A New Remedy for Bedtime Complaints of Worrying and Other Sleep-Disruptive Mental Activity. (Recall as noted above, SDI is a form cognitive shuffling.).
2. Colleagues at MacEwen University and I did another study on SDI the following year, this one seeing if it could be used for displaced sleep-phase disorder. That came out negative, and we weren’t surprised; the cognitive shuffle is not directly designed to address this condition. However, there are variants of SDKI which theoretically might be a bit helpful for that condition–but that remains to be seen. We did not publish this study.
3. Then, colleagues of mine at Université de Montréal and I ran another study involving the cognitive shuffle. This was presented at ICAP 2018 (International Conference on Applied Psychology, in Montreal) but there were no proceedings. The cognitive shuffle seemed to have had a positive impact, but as we mentioned in the conclusion of the poster a larger sample study is required. We did publish a World Sleep Congress abstract and presented a poster on a qualitative analysis of these data: Selham, Z., Guloy, S., Bastien, C., Beaudoin, L. P. & Carrier, J. (2019) “Research process and sleep app design lessons learned from the reflective examination of a sleep study”.

I hope these results and all the attention that the cognitive shuffle has received will encourage sleep labs around the world to test the somnolent information processing theory, to compare the various extant cognitive techniques, and to fine-tune existing cognitive techniques.

Are these cognitive techniques cures for insomnia

No, these techniques are not designed to be cures for insomnia. Insomnia is an entire syndrome. These cognitive techniques are at best an adjunct to CBT-I ( cognitive behavior therapy for insomnia) and sleep hygiene. Sleep hygiene is very important. If you drink a gallon of coffee in the evening, don’t expect the cognitive shuffle to come to your rescue.