Skip to content

Dataset · CognoSpeak

English speech records for early dementia detection

CognoSpeak/PROCESS-2

English spontaneous speech audio recordings and texts for cognitive impairment assessment.

What's inside

Contains English audio recordings and texts collected for remote cognitive diagnostics as a continuation of the PROCESS Challenge benchmark (ICASSP 2025). The dataset contains between 1,000 and 10,000 examples. The card specifies audio classification, text classification, and ASR tasks, with the file format unspecified.

What you can do with it

  • Training speech classification models to detect early signs of dementia
  • Testing ASR systems on patients' spontaneous speech
  • Creating multimodal classifiers using audio and transcripts
Tasks
Text classification, Automatic speech recognition, Audio classification
Languages
English
Size
1K–10K rows
Downloads this month
1.1K
Likes
48
Access
on request on Hugging Face

How to load it

from datasets import load_dataset

ds = load_dataset("CognoSpeak/PROCESS-2")