The Use of Machine Learning Methods in Neurodegenerative Disease Research: A Scoping Review
Bibliographic record
Abstract
Abstract Machine learning (ML) methods are increasingly used in clinical research, but their extent is complex and largely unknown in the field of neurodegenerative diseases (ND). This scoping review describes state-of-the-art ML in ND research using MEDLINE (PubMed), Embase (Ovid), Central (Cochrane), and Institute of Electrical and Electronics Engineers Xplore. Included articles, published between January 1, 2016, and December 31, 2020, used patient data on Alzheimer’s disease, multiple sclerosis, amyotrophic lateral sclerosis, Parkinson’s disease, or Huntington’s disease that employed ML methods during primary analysis. One reviewer screened citations for inclusion; 5 conducted data extraction. For each article, we abstracted the type of ND; publication year; sample size; ML algorithm data type; primary clinical goal (disease diagnosis/prognosis/prediction of treatment effect); and ML method type. Quantitative and qualitative syntheses of the results were conducted. After screening 4,471 citations and searching 1,677 full-text articles, 1,485 articles were included. The number of articles using ML methods in ND research increased from 172 in 2016 to 490 in 2020, with most of those in Alzheimer’s disease. The most common data type was imaging data (46.9% of articles), followed by functional (20.6%), clinical (14.2%), biospecimen (6.2%), genetic (5.9%), electrophysiological (5.1%), and molecular (1.1%). Overall, 68.5% of imaging data studies were in Alzheimer’s disease and 75.9% of functional data studies were in Parkinson’s disease. Disease diagnosis was the most common clinical aim in studies using ML methods (73.5%), followed by disease prognosis (21.4%) and prediction of treatment effect (13.5%). We extracted 2,734 ML methods, with support vector machine (n=651, 23.8%), random forest (n=310, 11.3%), and convolutional neural network (n=166, 6.1%) representing the majority. Finally, we identified 322 unique ML methods. There are opportunities for additional research using ML methods for disease prognosis and prediction of treatment effect. Addressing these utilization gaps will be important in future studies. Author Summary Few state-of-the-art scientific updates have been targeted for broader readerships without indulging in technical jargon. We have learned a lot from Judea Pearl on how to put things into context and make them clear. In this review paper, we identify machine learning methods used in the realm of neurodegenerative diseases and describe how the use of these methods can be enhanced in neurodegenerative disease research.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.036 | 0.145 |
| Meta-epidemiology (narrow) | 0.002 | 0.002 |
| Meta-epidemiology (broad) | 0.006 | 0.007 |
| Bibliometrics | 0.029 | 0.023 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.006 | 0.006 |
| Open science | 0.003 | 0.003 |
| Research integrity | 0.004 | 0.002 |
| Insufficient payload (model declined to judge) | 0.004 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".