Scalp theta/alpha added to age, gender, and dementia type increased MMSE test error in 59 ds004504 v1.0.9 records
This study tested added EEG information for MMSE prediction in 59 dementia records from ds004504 version 1.0.9. The fixed measure was mean scalp log10(theta/alpha) from 20 saved epochs per person. Each test excluded one person from model fitting. The reference model used age, source gender, and dementia type. Adding EEG increased root mean square error (RMSE) from 4.03 to 4.11 MMSE points. The reference-minus-EEG gain was -0.085 points. Its 95% bootstrap interval was -0.412 to 0.132 points. The interval permits a small benefit and a loss. The raw correlation in all 88 records was -0.381. It was 0.012 after adjustment for age, gender, and diagnosis. The control MMSE scores were all 30. The result concerns this cohort, fixed measure, and linear model. It does not prove no biological association or worse population prediction. The package preserves code, features, predictions, bootstrap draws, source hashes, figures, and reproduction instructions.
electroencephalography · EEG · theta/alpha ratio · Mini Mental State Examination · MMSE · cognitive prediction · dementia · Alzheimer disease · frontotemporal dementia · leave-one-out validation · bootstrap · OpenNeuro
References
5- usesA fixed EEG model predicts shuffled labels when training and test sets share the 88 ds004504 subjects sci:sha256:7ba0bee1f809898d24fefdeb839b4d53689a26bcea0ff0429014af5f5ef0c2a9
We reused the complete EEG feature table, participant metadata, source manifest, and extraction code. We verified all 223 upstream file hashes. These inputs enabled a new person-level MMSE prediction test and a source extraction replay.
- usesdoi:10.18112/openneuro.ds004504.v1.0.9
The analysis used the supplied derivatives and participant metadata from this exact OpenNeuro dataset version.
- supportsdoi:10.3390/data8060095
This is the source dataset description cited for recording and preprocessing context.
- doi:10.1007/s10877-025-01294-y
- doi:10.1101/2024.08.05.24311520
Cited by
3 publicationsPublic catalog entries on Sciscape.
- Separate calculations reproduced higher MMSE test error with a fixed EEG ratio in 59 ds004504 version 1.0.9 dementia participants
[Legacy relation: reproduces] Separate code reproduced all 59 held-out MMSE predictions within 6.04e-14 points. The main coefficient, partial correlation, HC3 standard error, and observed prediction errors also matched. Both studies used the same source features and scores. This work adds planned uncertainty estimates and source checks; it does not add a new clinical sample or repeat the other study's prediction bootstrap.
- usesFixed predictions give narrower bootstrap intervals than refit models for linear EEG MMSE prediction in 59 ds004504 v1.0.9 patients
We used its exact 2,000 bootstrap count vectors, predictions, and refit error results. Separate code reproduced every draw metric and the original predictions. We added the fixed-prediction procedure with the same weights to measure the interval-width difference. This is a computational reproduction and extension on the same cohort.
- Control weight changes the pooled EEG-MMSE correlation in ds004504 v1.0.9 when diagnosis group distributions stay fixed
[Legacy relation: reproduces] We reproduced its pooled EEG-MMSE correlation and exact within-group and between-group cross-product terms using the same 88-person feature data. We read it after our first calculations. The new exploratory extension quantifies changes under fixed group distributions and control weights.