Research collaboration

Instrumented for validation. Open to studies.

Neuropsych administers cognitive tasks online in English and Spanish, scores them deterministically against published normative tables, and records every attempt and retest in a research dataset that grows daily. The evidence base grows with use, but slowly. Dedicated studies of convergent validity, test-retest reliability and known-groups validity would establish it faster, and the infrastructure to run them is in place. We are looking for research groups to take them on.

Research infrastructurestatus
  1. Scoring engineDeterministic rules, published normative tables, every rule documented.
    In place
  2. Participant-level exportPseudonymized CSV or JSON per study, with trial-level data where logged.
    In place
  3. Open research datasetDe-identified first attempts, CC-BY 4.0 with a DOI, counted weekly.
    Accruing
  4. Retest captureRepeat administrations recorded with their interval.
    Accruing
  5. Validation studiesConvergent validity, test-retest reliability, known-groups validity.
    Open to partners

Filled: built and running. Ring: accruing from live use. Dashed: requires a research partner.

Three studies that would accelerate the evidence base

Each is within the scope of a single laboratory, or a doctoral project with ethics approval and participant access. The battery, the exports and the documentation are ready.

01

Convergent validity with established batteries

Do the online tasks measure the same constructs as their laboratory counterparts?

Design
N = 80 to 100 adults complete Neuropsych and an established battery (WAIS‑V subtests, D-KEFS or the Conners CPT-3) in counterbalanced order. Correlations between matched subtests.
Interpretation
Correlations of .60 or higher on matched subtests would support convergent validity.
02

Test-retest reliability

Are scores stable across repeated administration at a controlled interval?

Design
N ≥ 50 adults complete the battery twice, 2 to 4 weeks apart. Intraclass correlation coefficient per task. Live retests accrue without an interval protocol; a controlled study would settle it sooner.
Interpretation
ICC ≥ .80 is the conventional threshold for individual-level monitoring.
03

Known-groups validity in clinical samples

Do scores discriminate adults with a documented ADHD or dyscalculia diagnosis from controls?

Design
Three groups of N ≥ 30 adults: ADHD, dyscalculia and no diagnosis. ROC analysis on the relevant tasks.
Interpretation
AUC ≥ .80 would support use as a screening indicator for referral to formal assessment.

Can you run one of these?A protocol, a sample and ethics approval are what each study needs. We provide the rest.

Get in touch

What Neuropsych provides

Available now. Describe the protocol and we will state what can be supported.

The de-identified dataset is documented separately on the open research data page.

Administration
Participants complete the tasks online at no cost. Each task has a stable URL, so a protocol can administer any subset in a fixed order.
Participant-level data
With participant consent, pseudonymized results as CSV or JSON: one record per administration with scored metrics, trial counts and response-time summaries, and trial-level data where the task logs it.
Scoring documentation

Each scoring rule and the normative table it references are specified on the scoring engine page, so the procedure can be audited before data collection.

Scoring engine documentation
Ethical review
Ethical review remains with the investigator's institution. We supply task descriptions and data storage and retention details for the application.

A cognitive lab your students run themselves

Instructors already assign the tasks as lab exercises. Create a class link and your class gets a page of its own: every student gets their result, and you see the class data as it comes in.

See an example class summary
  1. Students open the linkEach completes the tasks you set and sees their own result straight away.
    Week 1
  2. You receive the class summaryPseudonymous rows with scores and percentiles, and a CSV for the stats session.
    Week 2
  3. The class reads its own distributionAgainst published population norms, with every scoring rule open to inspection.
    Week 2

Teaching a course this term?Create the class link yourself: name the class, pick the tasks, paste it in the learning platform.

Create a class link

Spanish normative data are the largest gap

Most tasks are administered in Spanish. None is referenced to Spanish norms yet.

The Spanish animal fluency tables of the NEURONORMA project (Casals-Coll et al., 2013) are for words said aloud, and on the site people type, so animal fluency is reported without a comparison in both languages. The other Spanish administrations are referenced to the English tables until Spanish normative data exist. The numerical tasks include dot and digit comparison, number line estimation, arithmetic fact retrieval, place value and number dictation.

Do you work with Spanish-speaking adults?A norming or validation study in Spain or Latin America would close the gap for every Spanish task at once.

Get in touch

Describe your study

A short note on your research area and the form of involvement you have in mind is sufficient. A study is not a prerequisite; methodological advice is also useful.

  • University researchers with ethics review access and participant pools, particularly for the convergent validity and norming studies.
  • Clinical neuropsychologists with access to adults holding a documented ADHD or dyscalculia diagnosis.
  • Doctoral students planning a validation or norming thesis.

Or write to hi@neuropsych.io.

Cristian Dominguez Rein-Loring, maintainer. Neurodivergent; background in creative technology.LinkedInWebsite

Message received

Thanks for reaching out. We will reply by email.