Audio Processing Updates Can Look Like Cognitive Change in Alzheimer’s Speech Data, Cephalgo CTAD 2026 Poster Shows
One simulated audio processing update moved 27% to 32% of people between MCI and Alzheimer's groups; recalibration
Press Release Disclaimer: This is a press release distributed through the XPR Media network. It has not been independently verified by our newsroom.
![]()
One simulated audio processing update moved 27% to 32% of people between MCI and Alzheimer’s groups; recalibration removed 95% to 97% of the false trend
BOSTON, MA, UNITED STATES, November 16, 2026 /EINPresswire.com/ — An audio processing update on a participant’s phone can change speech recordings enough to look like cognitive decline or improvement, according to a poster Cephalgo presents at the Clinical Trials on Alzheimer’s Disease conference (CTAD 2026) in Boston, 16-19 November. In Cephalgo’s own checks, a minor iOS release changed audio processing in measurement mode on both iPhones checked, and Android and web browser updates also changed audio processing settings. In simulated year-long trials with a known decline added, one audio processing update made speech timing features decline up to 39% faster or slower than the true decline.
Poster P446, “Noise-cancellation variability as a source of longitudinal drift in speech-based digital endpoints for Alzheimer’s disease trials,” used three published speech datasets. In the hardest comparison, Alzheimer’s disease versus MCI in one dataset, one simulated audio processing update moved 32% of people to the other group with standard handcrafted speech features, and 27% with Cephalgo’s Multi-stream model, although their condition had not changed. Multi-stream, which combines linguistic, acoustic and self-supervised analysis, stayed more stable across recording conditions than standard handcrafted features. After the worst audio processing update, its sensitivity and specificity were 17 to 19 percentage points higher. Recalibration from the audio removed 95% to 97% of the false trend, so an audio processing update no longer looked like decline or improvement.
“Alzheimer’s trials often run 18 months or longer, and participants’ mobile devices can update their software more than once in that time,” said Lisa Chiang, PhD, CEO of Cephalgo. “Sponsors need to know that a change in a speech reading reflects a change in the participant’s disease, not in the mobile device.”
Sponsors running speech endpoints can take four steps. Record the mobile device model and software version with every recording. Ask the vendor how its speech measures behave when audio processing changes. Test measures on the same audio processed in different ways. Before unblinding, state in the analysis plan how readings from before and after a change are compared.
Cephalgo estimates MMSE and MoCA scores from a few minutes of speech. MMSE was tested on a public challenge test set and MoCA in a separate published cohort. Each result is explained by named speech features, such as pauses, word choice and fluency. The analysis works on recordings already collected, including from running or completed studies.
The poster builds on Cephalgo’s robustness work in Parkinson’s disease and other motor studies at ISCTM 2026 and in Basel (doi:10.5281/zenodo.24060676, doi:10.5281/zenodo.24060707) and two ePosters at AD/PD 2026 (doi:10.5281/zenodo.24060705).
Poster P446 is on display Thursday 19 November and in the CTAD virtual poster hall. Poster details are at https://www.cephalgo.com/ctad-2026 and the full poster is available on request from Lisa.Chiang@cephalgo.com.
About Cephalgo
Cephalgo SAS, founded in 2021 and based in Strasbourg, France, is a speech biomarker company for CNS clinical trials in Alzheimer’s disease, Parkinson’s disease, Huntington’s disease and ALS. It is supported by a €2.5M European Innovation Council award. More at https://www.cephalgo.com
Lisa Chiang
Cephalgo SAS
+33 7 88 40 85 06
Legal Disclaimer:
EIN Presswire provides this news content “as is” without warranty of any kind. We do not accept any responsibility or liability
for the accuracy, content, images, videos, licenses, completeness, legality, or reliability of the information contained in this
article. If you have any complaints or copyright issues related to this article, kindly contact the author above.
![]()
Media gallery

