Predicting Conversion from SCD to MCI: A Machine Learning Study
Kamal, F.; Metz, A.; Chadwick, K.; Moqadam, R.; Morrison, C.; Dadar, M.. medRxiv (preprint), 2026-09-08
Topic: Risk, prevention and care in population studies
BACKGROUNDSubjective cognitive decline (SCD) may precede mild cognitive impairment (MCI), but not all individuals with SCD progress to MCI. Identifying which individuals are most likely to convert and over what time frame remains an important goal in Alzheimers disease research. MRI measures of white matter hyperintensity (WMH) burden and gray matter (GM) atrophy may improve prediction beyond demographic and cognitive predictors, but their incremental value across different time intervals has not been established. METHODSData were obtained from four longitudinal cohorts (ADNI, NACC, CIMA-Q, and PREVENT-AD). A total of 1,352 participants with SCD at baseline were included. Machine learning models (logistic regression, random forest, XGBoost) were used to predict conversion from SCD to MCI at 2 years, 3 years, 4 years, and 5-year horizons. Four feature sets were compared: base (age, sex, education, APOE4, hypertension), base and cognition (adding MoCA and Trail Making Test Part B), base and MRI (adding regional WMH and GM volumes), and combined (all features). RESULTSThe base and MRI set achieved the highest Area Under the Curve (AUC) at the 3-year (0.900), 4-year (0.879), and 5-year (0.923) horizons. The combined set achieved the highest AUC at the 2-year horizon only (0.884). MRI features produced larger AUC gains over the base model than cognitive features at the 3, 4, 5-year horizons. Parietal WMH was the most frequently selected MRI predictor. CONCLUSIONSMRI features, particularly regional WMH and GM volumes, provided greater predictive value than cognitive features at longer prediction horizons. A select number of regional MRI features predicted SCD to MCI conversion with high accuracy up to 5 years in advance.
Key findings
- The combined set achieved the highest AUC at the 2-year horizon only (0.884).
- MRI features produced larger AUC gains over the base model than cognitive features at the 3, 4, 5-year horizons.
- Parietal WMH was the most frequently selected MRI predictor.
- CONCLUSIONSMRI features, particularly regional WMH and GM volumes, provided greater predictive value than cognitive features at longer prediction horizons.
- A select number of regional MRI features predicted SCD to MCI conversion with high accuracy up to 5 years in advance.
Why it matters
This study used machine learning on MRI and other data to predict which people with subjective cognitive decline will develop mild cognitive impairment. MRI features, especially regional white matter hyperintensity and gray matter volumes, improved prediction more than cognitive tests at 3, 4, and 5 years, but this is a prediction tool, not a treatment or cure.
New to this subject? Start with our plain-English guide: Genetics →
Study details
- Study designlongitudinal study
- Participants1,352
- Evidence strengthModerate
How this study scored 16/100
Preprint — not yet peer-reviewed. It scored 35 on the same rubric, but non-peer-reviewed sources are capped at 16, so it shows as 16. Scored zero for effect reported and randomised / blinded because the record does not state it. We have no data on impact for its field and year, so that was left out: it earned 33 of the 94 points we could judge it on, which is 35 out of 100.
| Component | Points | Why |
|---|---|---|
| Journal tier | 2 / 10 | Tier 4 — not a peer-reviewed journal (medRxiv (preprint)) |
| Journal's citation record | 0 / 4 | Not a journal article, so the journal's record does not count |
| Study design | 7 / 12 | A longitudinal study following the same people over time |
| Strength of evidence | 3 / 5 | Rated “moderate” in our summary |
| Peer-review validation | 4 / 10 | Preprint — not yet peer-reviewed. Hosted on a preprint server. |
| How recent | 5 / 5 | 2026 — within 1 year(s) of the newest study here |
| Sample size | 3 / 4 | 1,352 participants |
| Effect reported | 0 / 3 | No effect size reported — scores zero (not stated in the record — scored 0) |
| Studied in people | 3 / 3 | Studied in people (participant count stated) |
| Randomised / blinded | 0 / 2 | No randomisation or blinding stated — scores zero (not stated in the record — scored 0) |
| Times cited | 0 / 7 | Cited 0 times (OpenAlex, 2026-10-10) |
| Citations per year | 0 / 5 | 0.0 citations a year over 2 year(s) since 2026 |
| Impact for its field and year | — / 6 | No field-weighted citation impact (OpenAlex has none yet) — left out of this score |
| Cited by studies here | 0 / 6 | No other study in this library cites it |
| Still being cited | 0 / 4 | Cited 0 times in 2024–2026 (2026 still in progress; OpenAlex) |
| Senior author's track record | 1 / 4 | Last-listed author Alzheimer’s Disease Neuroimaging Initiative: h-index 18 (OpenAlex, whole career) |
| Institution's track record | 1 / 3 | McGill University: institution h-index 1,373 (OpenAlex) |
| Multi-centre collaboration | 1 / 3 | OpenAlex lists 3 institution(s) in 1 country |
| Indexed in PubMed | 1 / 2 | Not in PubMed, but has a DOI |
| Free to read | 2 / 2 | Free to read |
| Points | 33 / 94 | Scored on 19 of 20 parts. We have no data for the rest, so they are left out: 33 of the 94 points we could judge it on is 35 out of 100. |
| Total | 35 / 100 | Capped to 16 because this is not a peer-reviewed article. |
What each of these 20 components measures, and what this score is not →
Read the original
Full text: free to read — OpenAlex lists this copy as free to read: the link above (checked 6 Feb 1020).
Cite this study APA, Vancouver, BibTeX · download .bib or .ris
Built from the record alzhub holds for this paper, not looked up. Volume, issue and page numbers are not held in the record. Check the citation against the original before you use it.
APA 7th
Kamal, F., Metz, A., Chadwick, K., Moqadam, R., Morrison, C., & Dadar, M. (2026). Predicting Conversion from SCD to MCI: A Machine Learning Study. medRxiv (preprint). https://doi.org/10.64898/2026.09.04.26362240
Vancouver
Kamal F, Metz A, Chadwick K, Moqadam R, Morrison C, Dadar M. Predicting Conversion from SCD to MCI: A Machine Learning Study. medRxiv (preprint). 2026. doi:10.64898/2026.09.04.26362240.
BibTeX
@article{kamal2026-13d655ad,
author = {Kamal, F. and Metz, A. and Chadwick, K. and Moqadam, R. and Morrison, C. and Dadar, M.},
title = {{Predicting Conversion from SCD to MCI: A Machine Learning Study}},
journal = {medRxiv (preprint)},
year = {2026},
doi = {10.64898/2026.09.04.26362240},
url = {https://doi.org/10.64898/2026.09.04.26362240}
}
Studies this one cites
Referenced by this study and also in the alzhub library, so you can read them here. Showing 12 of 13.