Neuropsychology is increasingly moving beyond paper-and-pencil tests toward digital assessment, immersive environments, remote rehabilitation and artificial intelligence. These technologies can provide richer information about how a person remembers, plans, pays attention, solves problems and responds to complex situations.
But more data does not automatically mean a better neuropsychological assessment.
A person may perform well on a structured computer task and still experience major difficulties organizing a meal, shopping independently, returning to work or managing multiple demands at home.
This creates one of the central questions of digital neuropsychology: How closely does performance inside a digital assessment reflect cognitive functioning in everyday life?
What is neuropsychology?
Clinical neuropsychology examines relationships between the brain, cognition, emotion, behaviour and everyday functioning.
Neuropsychological assessment may explore attention, processing speed, memory, working memory, executive functions, language, visuospatial abilities, social cognition, praxis, reasoning, and behavioural and emotional functioning.
Neuropsychologists may work with people affected by stroke, traumatic brain injury, acquired brain injury, neurodegenerative disorders, epilepsy, brain tumours, multiple sclerosis, Parkinson’s disease, developmental conditions, psychiatric disorders and other neurological or medical conditions affecting cognition.
From paper-and-pencil tests to digital neuropsychology
Traditional neuropsychological assessment has developed around carefully standardized tests administered under controlled conditions.
These tests remain extremely important because of their standardized administration, normative data, established psychometric properties and decades of clinical experience.
But traditional tests also have limitations. Real life is rarely as controlled as a quiet consultation room.
Daily activities often involve multiple simultaneous demands: remembering intentions, ignoring distractors, switching between tasks, managing time, making decisions, navigating environments and interacting with other people.
Digital technology creates opportunities to complement traditional assessment with tasks that reproduce some of this complexity.
Ecological validity: one of the biggest challenges in neuropsychology
Ecological validity concerns the extent to which an assessment meaningfully relates to behaviour and functioning outside the testing environment.
This is particularly important in neuropsychology.
Two people can obtain similar scores on a conventional planning test while functioning very differently in daily life.
Simply making a test look realistic does not automatically make it ecologically valid; the relationship between test performance and real-world functioning must actually be demonstrated.
Eight technologies changing neuropsychology
1. Computerized cognitive assessment
Computerized assessment can offer standardized presentation, precise reaction-time recording, automatic scoring, controlled stimulus timing, large numbers of trials and detailed error recording.
A computer may record differences that are difficult to capture manually, including response times, error patterns, hesitation, variability and learning curves.
However, digitizing a traditional test does not automatically create an equivalent test. Interface, screen size, input method, device latency and technology familiarity can influence performance.
2. Virtual reality for neuropsychological assessment
Virtual reality is particularly relevant because it can combine experimental control with more realistic situations.
A person may be asked to perform tasks inside a supermarket, kitchen, street, city, office, classroom or another everyday environment.
These scenarios can reproduce more complex cognitive demands than conventional tasks, but clinical impact and validation remain important issues.
3. Executive functions in immersive environments
A virtual task can require the participant to remember instructions, prioritize, inhibit irrelevant actions, switch between objectives, monitor time, correct mistakes and manage distractors.
This may reveal difficulties that are less apparent during highly structured testing.
4. Artificial intelligence and machine learning
Artificial intelligence may help identify patterns across large cognitive datasets.
Potential applications include classification, prediction of cognitive decline, combining several cognitive variables, identifying subtle patterns of errors, optimizing test batteries and supporting personalized monitoring.
AI-generated predictions should not be confused with clinical diagnoses.
A high-performing model may still be affected by biased datasets, demographic differences, small training samples, poor generalizability and unclear decision processes.
5. Digital cognitive rehabilitation
Digital platforms can provide structured cognitive exercises targeting attention, memory, working memory, processing speed, planning, inhibition and problem solving.
Technology can increase practice opportunities and automatically modify difficulty.
But cognitive rehabilitation should not become an endless sequence of abstract exercises.
Improving performance on a memory game does not necessarily mean the patient will remember appointments or medication.
6. Cognitive telerehabilitation
Remote platforms can allow clinicians to prescribe activities, review performance, conduct video sessions, coach compensatory strategies, involve family members and monitor progress.
A hybrid model may be particularly relevant: in-person assessment → supervised intervention → structured home practice → remote follow-up → reassessment.
7. Ecological momentary assessment and digital monitoring
Instead of evaluating cognition during one appointment, digital tools can collect information repeatedly across time.
A smartphone may ask brief questions at different moments during the day about fatigue, attention, mood, cognitive complaints, environmental context and day-to-day variability.
These approaches can be promising complementary tools but introduce major issues involving privacy, missing data, participant burden, interpretation and consent.
8. Extended reality: VR, AR and MR
Extended reality includes virtual reality, augmented reality and mixed reality.
Most current research focuses on VR, while AR and MR may eventually allow cognitive tasks to remain connected to the real physical environment.
The virtual supermarket as a neuropsychological task
A supermarket is an excellent example of why immersive assessment can be interesting.
The person may need to remember a list, navigate, search visually, ignore irrelevant products, compare information, manage time, update working memory, respond to unexpected events, follow rules and monitor completed items.
A virtual supermarket can potentially record total duration, correct products, incorrect selections, items picked up and replaced, number of list consultations, route through the environment, revisited areas, pauses, sequence of actions and rule violations.
This creates a richer cognitive behaviour profile.
Errors may be more informative than final scores
Digital assessment creates an important opportunity: analysing how the person fails, not only whether they succeed.
Two patients can complete a task in the same amount of time but demonstrate very different behaviours.
One may follow an efficient route, make one error and correct it. Another may repeatedly revisit the same areas, forget completed steps, select distractors and fail to recognize errors.
The final duration is identical. The cognitive behaviour is very different.
From scores to cognitive behaviour
| Digital variable | Potential interpretation | Important limitation |
|---|---|---|
| Completion time | Processing efficiency | Speed may trade off with accuracy |
| Number of errors | Task performance | Error type matters |
| Repeated errors | Monitoring or memory difficulty | Requires contextual interpretation |
| Route | Planning and search strategy | Efficient route depends on task |
| Pauses | Hesitation or processing | Could reflect interface difficulty |
| Distractor selections | Inhibition/attention | May also reflect misunderstanding |
| List consultations | Memory strategy | Consulting the list may be adaptive |
| Self-correction | Error monitoring | Requires clear event definition |
| Reaction time | Processing speed | Device and motor factors matter |
Four realistic clinical scenarios
Acquired brain injury: executive functioning
A patient with acquired brain injury may achieve relatively good scores on traditional tasks but report major difficulty managing activities at home.
A virtual shopping or cooking task could add information about planning, sequencing, prospective memory, inhibition, error monitoring and distractibility.
VR should complement—not automatically replace—the established assessment battery.
Stroke: cognitive rehabilitation
A person after stroke may work on attention and executive functioning through progressively more complex virtual tasks.
The clinician may begin with one objective, few distractors and unlimited time, then add multiple objectives, distractors, time constraints and memory requirements.
The objective is not merely to become better at the simulation. Strategies developed in the task should progressively transfer to real activities.
Mild cognitive impairment
A person presenting subtle cognitive complaints may complete digital tasks that capture not only accuracy but also reaction time, variability, learning, navigation and error patterns.
AI may eventually help identify multidimensional patterns that deserve further clinical investigation.
But an algorithmic risk score should not be presented as a diagnosis without appropriate clinical assessment.
Social cognition
Virtual environments can also create controlled social situations.
A patient might need to interpret another person’s intention, identify emotional information, respond appropriately and manage conflicting contextual cues.
Virtual social behaviour still needs to be validated against meaningful real-world social functioning.
Neuropsychological rehabilitation in everyday environments
Digital tools become especially interesting when assessment and rehabilitation can be linked.
An assessment may show that a patient forgets instructions when distracted, repeatedly returns to completed tasks or fails to monitor time.
Rehabilitation can then target those exact processes through external reminders, checklists, verbal self-instruction, environmental simplification, structured scanning and pause-and-check strategies.
The same virtual environment can then be reused with altered conditions to observe whether the strategy works.
This creates a loop: assessment → strategy → practice → adaptation → reassessment.
Can neuropsychologists and therapists share the same virtual environment?
Potentially, yes.
Shared virtual environments can allow the clinician to be present in the same scenario as the patient.
The clinician might demonstrate a strategy, provide cues, deliberately create a distractor, observe the response and progressively reduce assistance.
This is particularly interesting when cognitive rehabilitation needs to move from highly structured exercises toward functional tasks.
AI-generated cognitive exercises
Generative AI can help clinicians prepare category exercises, memory material, problem-solving scenarios, reading material, sequencing activities, patient education and compensatory strategy worksheets.
This can reduce preparation time.
But AI-generated cognitive content needs careful professional review for difficulty, language, cultural appropriateness, ambiguity, factual accuracy and alignment with the cognitive target.
Digital biomarkers and cognitive phenotyping
Digital systems can potentially generate large numbers of behavioural variables.
When repeated over time, these may create individual profiles including reaction-time variability, navigation patterns, error frequency, pauses, response consistency and activity patterns.
These features are sometimes discussed as potential digital biomarkers or components of digital phenotyping.
This field is promising but requires substantial caution. A behavioural pattern collected from a device does not automatically indicate a neurological disorder.
A major risk: over-measurement
Digital systems make it possible to collect hundreds of variables.
That does not mean we should.
Too many metrics create false positives, difficult interpretation, statistical noise, clinician overload and increased privacy burden.
The relevant question is not: How much data can we collect?
It is: Which variables help us understand this person’s cognitive functioning or make a better rehabilitation decision?
What does current evidence tell us?
Digital neuropsychology is promising, but it is not one homogeneous field.
Immersive VR can reproduce more complex cognitive demands than many conventional tasks and may help differentiate some clinical populations from healthy controls.
However, validation remains a major issue. Many tools still lack sufficient normative, reliability or clinical validation data.
Digital tools should therefore currently be viewed primarily as validated instruments when evidence exists + complementary assessment tools + rehabilitation environments + research platforms.
They should not be assumed to replace standardized neuropsychological batteries simply because they appear more realistic.
Nine questions before adopting a digital neuropsychology tool
- What cognitive process or functional difficulty are we trying to understand?
- Has the tool been validated for this purpose and population?
- Are normative data available when they are required?
- Does it add information that established testing does not provide?
- Are the digital metrics clinically interpretable?
- Could technology familiarity influence performance?
- What happens if cybersickness, fatigue or sensory discomfort occurs?
- How does performance relate to real-world function?
- Will collecting this additional data actually influence clinical decisions?
The future: from cognitive scores to cognitive behaviour
Traditional neuropsychology has developed extremely sophisticated ways of measuring cognition.
Digital technology should not discard that foundation.
Its strongest contribution may be to add another level: understanding how cognition unfolds over time inside complex activities.
Instead of knowing only that the patient obtained a certain score, clinicians may increasingly examine the sequence of decisions, types of errors, response to distractors, self-correction, use of strategies, navigation and behaviour under time pressure.
At Remotion, this principle is particularly relevant to immersive daily-life scenarios.
A virtual supermarket, kitchen or other functional environment can generate more than an end score.
It can potentially help clinicians observe how a person approaches, organizes and adapts to a complex task.
The goal is not to replace neuropsychological assessment. It is to create additional bridges between standardized cognition and the cognitive demands of everyday life.
Frequently asked questions
Can VR replace traditional neuropsychological tests?
Not currently as a general rule. Some VR tools show promising validity, but many still lack sufficient normative, reliability or clinical validation data. They are often better considered complementary tools.
Why is ecological validity important?
A cognitive test is clinically useful partly because it helps us understand functioning beyond the test itself. Ecological validity examines this relationship with everyday behaviour and functioning.
Can AI diagnose dementia from cognitive tests?
AI can identify patterns associated with cognitive impairment and may support detection or prediction, but algorithmic output must be integrated with appropriate clinical assessment and validated diagnostic processes.
Can virtual environments assess executive functions?
Yes, they can create situations involving planning, inhibition, switching, memory and multitasking. Validation varies between tools.
Can cognitive rehabilitation be performed remotely?
In some populations, yes. Cognitive telerehabilitation is an expanding area, but evidence and suitability vary according to condition, intervention and individual needs.
Are more digital metrics always better?
No. Collecting variables without a clear clinical purpose may make interpretation harder rather than better.
Is a virtual supermarket a neuropsychological test?
It can be developed as one, but simply creating a supermarket simulation does not make it a validated neuropsychological instrument. Clinical use as a test requires appropriate psychometric evidence.
Selected references and further reading
- Pieri L, Tosi G, Romano D. Virtual reality technology in neuropsychological testing: A systematic review. Journal of Neuropsychology, 2023.
- Cavallo ND et al. Immersive Virtual Reality in Neuropsychological Assessment of Acquired Brain Injury: a Systematic Review and Meta-Analysis. Neuropsychology Review, 2026.
- Faria AL et al. Ecologically valid virtual reality-based technologies for assessment and rehabilitation of acquired brain injury: a systematic review. 2023.
- Veneziani I et al. Applications of Artificial Intelligence in the Neuropsychological Assessment of Dementia: A Systematic Review. Journal of Personalized Medicine, 2024.
- Pinto JO et al. Ecological validity in neurocognitive assessment: Systematized review, content analysis, and proposal of an instrument. Applied Neuropsychology: Adult, 2025.
- Despoti A et al. Effectiveness of remote neuropsychological interventions: A systematic review. 2024.
- A systematic review of cognitive telerehabilitation in patients with cognitive dysfunction.
- Extended reality for neurocognitive assessment: A systematic review. 2025.
- Exploring Cognitive Functions Through Virtual Reality: a Systematic Review of Ecologically Valid Assessments. 2026.
- Current limitations in technology-based cognitive assessment for severe mental illnesses: a focus on feasibility, reliability, and ecological validity. 2025.