A pragmatic clinical trial conducted within routine psychogeriatric rehabilitation examined how progressively introduced extended reality could be integrated and adapted within routine cognitive stimulation for older adults with mild or early cognitive impairment, with exploratory comparison of cognitive and functional change against computerised cognitive training alone. Carried out at the psychogeriatric day hospital of the Centre d’Atenció Intermèdia El Carme in Badalona, Spain, the study evaluated 70 structured rehabilitation courses undertaken by 59 participants undergoing twelve weeks of clinician-supervised cognitive intervention. Across 45 treatment courses incorporating extended reality into regular care, clinicians recorded 729 distinct participant-exposures, of which 715, or 98.1 percent, were delivered in full without premature session termination. Beyond establishing feasibility in a real-world healthcare environment, exploratory statistical comparisons against participants receiving conventional computerized cognitive training alone indicated adjusted between-group differences for global cognitive screening metrics, which the authors state warrant further controlled study but do not establish an XR-specific treatment effect. However, the researchers emphasized that immersive technology cannot be treated as a monolithic medical intervention, as operational success and cognitive burden varied dramatically depending on whether participants engaged through passive visual exploration, direct natural hand tracking, or complex handheld hardware controllers.

The therapeutic architecture was integrated directly into an existing non-pharmacological stimulation service rather than operating in an artificial research vacuum, creating a realistic pragmatic testbed for modern neurorehabilitation. Both the experimental and comparison cohorts received regular supervised computerized training utilizing the web-based NeuronUP platform, engaging in standardized modules targeting sustained attention, working memory, processing speed, and executive planning through tasks such as Bottle Caps, Making Words, and Find Your Luggage. In the extended reality arm, clinicians allocated approximately fifteen to twenty minutes of each scheduled 45-minute therapeutic session to immersive digital activities across two days per week, leaving the total session duration unchanged while altering the density and multimodal nature of cognitive stimulation. Group allocation was dictated by pre-existing service schedules and logistical cohort progression across rolling admissions rather than randomized allocation concealment, with the same specialized clinical team managing both arms. Participants averaged over 72 years of age, presenting diverse underlying etiologies ranging from Alzheimer disease and vascular cognitive impairment to psychiatric conditions and neurodegenerative disorders, reflecting the true diagnostic heterogeneity encountered in day hospital psychogeriatric practice.

Rather than thrusting participants directly into demanding interactive simulations, the clinical team instituted a five-stage progressive familiarisation model that prioritized incremental interaction complexity over sheer sensory immersion. The initial phase deployed Benestar VR, an individual seated experience offering passive visual immersion and audio guidance in tranquil settings to allow clinicians to observe tolerance, head orientation, and comfort without requiring task performance. The second stage introduced Reality Telling 360-degree video, an environment delivered in small groups under shared therapist supervision, where participants visually explored familiar or recognisable locations while discussing memories without manipulating virtual objects. The third stage transitioned into REDOCVR passthrough mixed reality, blending digital cognitive-motor puzzles and reaching activities into the physical room while participants maintained full visual awareness of their real surroundings using direct hand tracking. The fourth stage introduced fully immersive everyday-task virtual reality, challenging individuals with complex instrumental activities of daily living such as grocery shopping, exact cash payments, and sequential recipe preparation. Only in the fifth stage did participants encounter Enhance VR, a commercially available suite of structured cognitive training exercises that necessitated navigating abstract environments and utilizing physical handheld controllers.

Task understanding or execution difficulties were documented in mixed-reality cognitive exercises and immersive everyday-task virtual reality, while the introduction of handheld controllers during Enhance VR was associated with the highest frequency of such difficulties and the only documented interaction or interface problems. Task understanding or execution difficulties occurred in 51 of 273 participant-exposures within Enhance VR, affecting 25 of the 43 courses exposed to that software family, representing a 58.1 percent course-level rate of documented difficulty. By contrast, direct natural hand tracking in mixed reality and everyday functional simulations generated task execution difficulties in only 12.5 percent and 14.0 percent of courses respectively, despite demanding high-level executive sequencing, numerical calculations, and spatial memory.

Adverse symptoms were uncommon at the exposure level across the series of exposures, with no serious adverse events or symptom-related discontinuations documented. Out of 729 total participant-exposures, adverse symptoms were recorded in only 30 sessions, representing an overall exposure-level incidence of 4.1 percent and affecting 14 of the 45 participants across their rehabilitation courses. Mild transient dizziness constituted the most frequent complaint, documented across 19 exposures in nine individuals, followed by localized headset pressure, mild headache, and isolated post-session disorientation. Notably, every documented instance of dizziness during 360-degree video viewing occurred during recordings depicting dynamic camera movement through environments, whereas static, stationary scenes produced zero dizziness.

Patient-reported and clinician-reported usability assessments were completed using study-specific instruments adapted from the System Usability Scale, providing descriptive evaluations of the technology when embedded within supportive clinical scaffolding. Thirty-eight participants who completed full pre-post extended reality courses evaluated the programme using study-specific instruments adapted from the System Usability Scale, yielding an average patient-perceived usability score of 68.50 out of 100 and a high headset comfort and tolerability rating of 81.99 out of 100. Four attending professionals evaluating the intervention returned a composite usability score of 85 out of 100, with all four strongly agreeing that the programme could be integrated into routine practice. Qualitative feedback from therapists emphasized that guided 360-degree environments served as useful for facilitating spontaneous communication and shared discussion, while the immersive grocery shopping application was considered a functionally relevant activity.

Exploratory inferential analyses of primary cognitive outcomes indicated statistically significant adjusted differences favoring the extended reality condition across global cognitive screening metrics. Utilizing generalized estimating equations that accounted for within-participant clustering and controlled for baseline performance levels, the researchers observed an adjusted between-group difference of 1.62 points on the Mini-Mental State Examination and 1.88 points on the Montreal Cognitive Assessment. Both screening differences retained formal statistical significance following rigorous Holm multiplicity adjustment across the battery of eight clinical outcomes, yielding adjusted p-values of 0.013 for the Mini-Mental State Examination and 0.030 for the Montreal Cognitive Assessment. Detailed examination of the raw trajectories revealed that these between-group advantages were primarily driven by cognitive stability and slight gains among extended reality participants alongside mild performance declines within the non-extended reality control cohort over the twelve-week interval. Specifically, Mini-Mental State Examination averages shifted from 25.84 to 26.00 in the extended reality arm compared to a drop from 26.39 to 24.87 in controls, while Montreal Cognitive Assessment scores rose from 18.55 to 19.37 versus a decline from 18.17 to 17.13.

In contrast to global screening stability, the intervention did not produce measurable transfer to standardized functional independence or secondary cognitive domains. Changes on the Lawton-Brody Instrumental Activities of Daily Living Scale showed an adjusted between-group difference of only 0.40 points, which was not statistically significant, reflecting known ceiling effects of the instrument in relatively independent day hospital attendees whose baseline capabilities left minimal margin for measurable numerical gains. Visual attention and working memory tests, including the Trail Making Test Part A, Digit Span Forward, Digit Span Backward, and the Symbol Digit Modalities Test, similarly showed no significant adjusted divergence between arms. Analysis of Trail Making Test Part B provided intriguing qualitative patterns regarding executive task completion: the proportion of participants able to successfully complete the alternating sequence task rose from 60.5 percent to 78.9 percent in the extended reality group while dropping from 82.6 percent to 65.2 percent in controls. Although the adjusted odds ratio for follow-up completion reached 3.99 favoring the experimental group, it remained nominally non-significant, and an initial completion speed advantage of 63.49 seconds among paired completers lost statistical significance following Holm correction.

The study authors maintain a rigorous and cautious stance regarding clinical interpretation, noting that pragmatic trial constraints prevent attributing cognitive differences exclusively to extended reality hardware or virtual environments. Because treatment allocation followed existing service grouping rather than individual randomization, unmeasured confounding, subtle variations in therapist attention, and novelty effects cannot be entirely ruled out. Furthermore, eleven individuals contributed sequential courses by participating in the control arm before crossing over to the extended reality condition without an intervening washout period, introducing potential carryover effects that required specialized clustered modeling. Sensitivity analyses restricting the dataset strictly to the first chronological course per participant confirmed positive nominal trends for global cognition and task completion speed, though statistical significance attenuated below the threshold of multiplicity correction in that reduced sample. Rather than claiming definitive therapeutic efficacy, the researchers conclude that clinician-supervised extended reality represents a viable, highly tolerable adjunct that enriches routine cognitive stimulation, provided that clinicians dynamically select software configurations that match the sensory, physical, and cognitive tolerances of older adults.

Subject of Research: Progressive extended reality integration in cognitive stimulation for mild cognitive impairment

Article Title: Progressive task-oriented extended reality in a cognitive stimulation programme for older adults with mild or early cognitive impairment: a pragmatic clinical trial of patient adaptation, usability and cognitive-functional outcomes

Article References: Ferrer Costa, J., Ciudad, M. J., Patricio, V., González Abad, M., Rodríguez García, J. L., & Bayona-Huguet, X. (2026). Progressive task-oriented extended reality in a cognitive stimulation programme for older adults with mild or early cognitive impairment: a pragmatic clinical trial of patient adaptation, usability and cognitive-functional outcomes. Frontiers in Virtual Reality, 7, Article 1995712. https://doi.org/10.3389/frvir.2026.1995712

Image Credits: AI Generated

DOI: 10.3389/frvir.2026.1995712

Keywords: extended reality, cognitive stimulation, mild cognitive impairment, neurorehabilitation, virtual reality, mixed reality, usability, clinical trial, Progressive, task-oriented, extended, reality