Recalla home Evidence Hub

What's established, what we're testing, and what's still open.

A transparency page, not a marketing page. It separates what published science has shown about speech and cognition from what Recalla has actually built and hypothesized — and it's honest about what hasn't been proven yet.

Prototype-stage · not a diagnostic device · synthetic examples only
External evidencePeer-reviewed science by the wider research community. Cited, with DOIs. Not Recalla's own work.
Company evidenceRecalla's own hypotheses, prototype, and results — dated, with honest empty states where results don't exist yet.
Published scienceWhat the literature has established about speech, language, and cognition.
External
Our hypothesisWhat Recalla proposes to test — framed as a hypothesis, not a claim.
Company
What we've builtThe prototype, pipeline, and framework that exist today.
Company
What we're validatingThe pilots and studies underway or planned to test the hypothesis.
Company
Future studiesThe longer-term validation roadmap toward clinical evidence.
Company
External evidence · published research

Scientific foundation

The idea that speech carries signals about cognition is not ours — it's a decades-old, peer-reviewed body of work. Here's what that literature actually establishes, including where it's cautious.

Speech and language change measurably in cognitive decline

Fluency, lexical diversity, pause patterns, and semantic content shift in neurodegenerative disease — and those shifts can be measured automatically.

Signals appear early — even at the MCI stage

Speech-derived measures help distinguish mild cognitive impairment from healthy ageing. A 2025 meta-analysis of 54 studies found ~80% pooled accuracy for MCI — with wide confidence intervals.

Language can shift years before diagnosis

Early-life linguistic ability predicted late-life cognition in the Nun Study; language features predicted future Alzheimer's onset in the Framingham cohort.

Connected speech tracks change over time

In autopsy-confirmed Alzheimer's, connected-speech features changed in step with disease progression — evidence that speech can be a longitudinal marker, not only a one-time classifier.

The field has standard methods and shared benchmarks

Standardized acoustic feature sets and open shared-task benchmarks mean methods can be compared rigorously rather than in isolation.

What this evidence does not yet show

Most published results are cross-sectional group classification (people with a diagnosis vs. without), not within-person tracking over time. Performance varies with task, language, and recording quality, and larger prospective validation is repeatedly called for.

External evidence · current practice

Current clinical landscape

Why a complementary signal could matter — and where a digital approach is being explored in the literature.

How cognition is assessed today

  • Brief cognitive tests (MMSE, MoCA) administered in clinic
  • Clinical interview and informant history
  • Neuropsychological batteries, usually at annual or longer intervals
  • Confirmatory biomarkers (imaging, CSF) where indicated

Documented limitations

  • Episodic — snapshots months or years apart
  • Rater-dependent and subject to practice / ceiling effects
  • Resource-intensive and often accessed late
  • Poorly suited to catching a person's earliest deviation from their own normal
57.4M → 152.8M
Estimated people living with dementia globally, 2019 → projected 2050 (Global Burden of Disease 2019).
DOI: 10.1016/S2468-2667(21)00249-8

Where digital, remote approaches are being explored

  • Digital biomarkers from everyday devices as an early-signal opportunity Kourtis 2019
  • Automated remote speech assessment detecting early impairment Robin 2021
  • Smartphone speech screening against amyloid-confirmed cohorts Fristed 2022
Company evidence · Recalla's position

Our scientific hypothesis

This is the line where external science ends and our own contribution begins. We state it as a hypothesis under evaluation — deliberately, because that's what it is.

The question we're testing
We hypothesize that frequent, low-friction voice check-ins can produce longitudinal signals that complement — not replace — episodic clinical assessment, and that a person's change from their own baseline is more informative than any single snapshot.

Notice what this is not: a claim that Recalla detects or diagnoses dementia. Whether these signals are clinically useful in real-world settings remains under prospective evaluation. The sections below are honest about how far along that evaluation is.

Company evidence · as of Q3 2026

What we've built & validated

The honest current state. Where results don't exist yet, we say so rather than implying them.

Built

Voice check-in pipelinePrototype capture → acoustic + linguistic feature extraction.
Acoustic feature layerBuilt on standard tooling (Praat via Parselmouth) and published feature sets.
Illustrative scoring frameworkPer-family signals combined into an explainable, illustrative composite.
Explainability by designEvery score decomposes into the signals that drove it — no black box.
Open technical previewPublic framework repository and an interactive, synthetic-data demo.

In progress

Concierge pilot designA small, human-in-the-loop feasibility pilot measuring adherence and clinician usefulness.
Design-partner conversationsDiscovery with caregivers and clinicians to pressure-test the workflow.
Task battery refinementDesigning check-in prompts with known content for consistent measurement.

Not yet — planned

Prospective pilot dataNo pilot results to report yet.
Clinical validationNo clinical validation has been performed.
Peer-reviewed publicationsNone yet; this section will list them as they appear.
Regulatory clearanceNot sought; Recalla is not a medical device today.
Recalla is a research-stage tool. It is not a diagnostic device, has not been validated for clinical use, and its scores are illustrative. The interactive demo runs on synthetic data; live check-in features compute on-device and produce modeled, non-diagnostic values.
Company evidence · plan, not results

Evidence roadmap

How we intend to move from feasibility to clinical evidence. These are planned milestones and directions — not commitments or completed results.

2026

Feasibility

Concierge pilot; measure adherence and clinician usefulness.

2027

Pilot study

Structured pilot; first longitudinal data readouts.

2028

Clinical validation

Prospective study against reference measures.

2029

Real-world eval

Multi-site evaluation in care settings.

2030

Longitudinal study

Large longitudinal cohort; peer-reviewed publication.

Dates are directional and will move as evidence accrues. We'd rather revise this roadmap in public than overpromise on it.
External evidence · reference library

Research library

Every reference below was verified to its source — real title, authors, journal, year, and DOI. It's meant to be useful to any reader working on voice and cognition, not only to us. Filter by topic.

Frequently asked

Scientific questions, answered plainly

No. Recalla is not a diagnostic device and does not detect, diagnose, or rule out any condition. It's designed to produce longitudinal signals that could complement clinical assessment. Any score it shows is illustrative and non-diagnostic.

The foundation is. Decades of peer-reviewed work show that speech and language change with cognition (see the Scientific foundation and Research library). What's ours — and still a hypothesis under evaluation — is the specific bet that frequent, low-friction, longitudinal voice check-ins are clinically useful in the real world.

To date: a working prototype, a feature-extraction pipeline, an explainable scoring framework, and an open technical preview. We have not completed a pilot, performed clinical validation, or published peer-reviewed results. A concierge feasibility pilot is in design. The What we've built section tracks this honestly.

Most prior work is cross-sectional group classification — telling apart people who already have a diagnosis from those who don't, at a single point in time. Recalla focuses instead on within-person change over time, as a complement to episodic testing. Whether that's more useful in practice is exactly what we intend to test.

The interactive demo uses synthetic participants only — no real patient data. Its live check-in features run in your browser and produce modeled, illustrative values from your own voice; nothing is uploaded. Pilots will be run under appropriate consent and ethics review.

Yes, by design. Every composite decomposes into the individual signal families that produced it, so a clinician can see why a value moved. Explainability is increasingly recognized as essential for clinical adoption and regulation of speech-based tools (Shankar 2025).

The full, DOI-linked Research library is on this page. Our public technical framework and the interactive demo are linked from the Resources page. This Evidence Hub will grow as our own results and publications appear.