Study: A digital cognitive stimulation program was tested for safety and effectiveness in adults with insomnia

What the study tested

Insomnia is a large and largely undertreated burden. The authors cite estimates that around 1.2 billion people worldwide meet insomnia criteria, with roughly 4.7 million diagnosed in Spain among a population of 47 million.¹ They note that although cognitive behavioral therapy for insomnia (CBT-I) is the recommended first-choice treatment, only an estimated 1% of diagnosed patients receive it,¹ while a meta-analysis confirms CBT efficacy but documents persistent reliance on medication [3].^[1]

Against that gap, this trial asked whether a computerized cognitive training program, framed as a digital therapeutic targeting the attentional control and inhibitory processes implicated in insomnia, could be delivered safely and was associated with changes in sleep and emotional state. Registered as NCT05050292, it reported combined Phase I and Phase II results.¹ Participants were recruited from the Sleep Unit of the Hospital Universitario de la Ribera in Spain: patients aged 25 to 55, diagnosed with insomnia disorder (307.42 [F51.01]), with sleep difficulties at least three times per week for at least three months. From 51 eligible candidates, 24 enrolled and 12 completed (mean age 44 years; 9 men and 3 women; all living in a rural area).¹

The intervention was a home-based computerized cognitive stimulation program (the CogniFit program) installed on participants' own smartphones. It began with a cognitive evaluation using the Cognitive Assessment Battery (CAB PRO), which built an individual profile that an Individualized Training System used to tailor the gamified activities; the tasks were designed to stimulate executive functions with a focus on inhibition skills.¹ In Phase II, participants trained on alternate days for 15 days, completing 7 sessions of about 45 minutes (5 hours 15 minutes total), each comprising 9 cognitively demanding games of about 5 minutes.¹

The main sleep outcome measures were the Insomnia Severity Index (ISI) and the Pittsburgh Sleep Quality Index (PSQI). Secondary measures of cognitive and emotional state were the Behavior Rating Inventory of Executive Function–Adult version (BRIEF-A), the Beck Depression Inventory-II (BDI-II), the State-Trait Anxiety Inventory (STAI), and the Penn State Worry Questionnaire (PSWQ), administered before (day 1) and after (day 15).¹ A preceding Phase I used a dose-escalation 3+3 design to set the maximum tolerated session length, with a structured adverse-event interview for safety.¹

What it found

Scores (day 1 vs day 15) were compared with paired t-tests. Four of the six instruments improved significantly, each with a large effect size; two did not.¹

On the emotional measures, depressive symptoms on the BDI-II fell from 39.75 to 33.58 (mean difference −6.17; t(11) = 2.87; p = 0.015; d = 0.83), and worry on the Penn State Worry Questionnaire decreased from 41.17 to 37.0 (mean difference −4.17; t(11) = 3.53; p = 0.005; d = 1.02).¹ The largest change appeared on the STAI-Trait anxiety scale, where scores dropped from 34.25 to 16.0 (mean difference −18.25; t(11) = 12.18; p < 0.001; d = 3.52).¹

Two outcomes did not reach significance. Executive functioning on the BRIEF-A moved from 125.33 to 119.0 but the change was not significant (t(11) = 1.43; p = 0.182; d = 0.41).¹ Because normality was violated for the ISI, a Wilcoxon test was applied and indicated no difference in insomnia-severity indices before versus after (p = 0.364; rank-biserial 0.36).¹

On safety, the Phase I cohort reported extreme fatigue (8 or above on a 1–10 scale) only after roughly 60 minutes of training, so the maximum tolerated dose was set at 45 minutes per session.¹ In Phase II, no session produced extreme fatigue or undesirable side effects, and the authors reported that improvements occurred "in the full absence of safety issues."¹ Adherence was reported as acceptable, against a 50% drop-out benchmark from trials of smartphone apps for depressive symptoms [6].^[1]

Limitations, what this study does not show

The authors were explicit that this was an early-stage trial that "closely resembles a case study approach." They named the small sample size (12 completers), the lack of a control group, and the unknown future outcome as factors limiting generalizability.¹ A further limitation was the clinicians' daily monitoring: because a Phase I/II study requires that supervision, it remained unknown whether participants would achieve comparable results in an unsupervised, self-administered setting.¹

The design also leaves the durability question open. Of the 24 enrolled participants, 12 withdrew, a 50% dropout rate, partly because the protocol was too time-demanding.¹ The single-arm, pre-post structure means improvements cannot be separated from expectancy, supervision, or natural fluctuation. The authors noted that some questionnaires (the ISI and BRIEF-A) ask about behavior "during the past month," so they may not be sensitive to a 15-day intervention, one reason those two measures did not move.¹ Long-term effectiveness was, in their words, "still unknown," and they called for randomized trials with longer follow-up.¹ This trial reports feasibility and safety signals in a small sample, not proof that the program treats insomnia.^[1]

Where this fits in the broader evidence

The authors described this as, to the best of their knowledge, "the first attempt to obtain quantitative data regarding the safety and potential efficacy of a CCT in insomnia."¹ Its sleep-quality signal is consistent with an earlier randomized trial of the same personalized training in older adults with insomnia [2], summarized on our page on cognitive training and sleep quality in older adults. It sits in a field where European guidelines position CBT-I as first-line and pharmacology as an adjunct [4], and digital therapeutics are advancing through clinical pipelines [5]. See also the guide on cognitive training for insomnia, the overview of brain training in older adults, and the research index.