85 Predicting Conversion to Mild Cognitive Impairment in Parkinson’s Disease: a Random Forest Machine Learning Model Based on Parkinson’s Progression Markers Initiative Dataset.
Bibliographic record
Abstract
Objective: Mild cognitive impairment (MCI) is common in Parkinson’s disease (PD). Recent scientific advances show that MCI in PD could also be impacted by neuropsychiatric symptoms (such as apathy, anxiety, depression), dopaminergic deficiency (more striatal denervation associated with MCI) and certain genotypes such as in APOE E4, MAPT H1 or SNCA C/C carriers. We used a python-based random forest machine-learning algorithm (scikit-learn) in order to evaluate the factors that are mostly involved in the MCI conversion over a 5-year follow-up period. Participants and Methods: Baseline data of healthy individuals and participants with Parkinson’s disease were extracted from the PPMI dataset. All participants also had the evaluations of their cognitive status, neuropsychiatric symptoms (hallucinations, anxiety, apathy, depression, sleepiness, impulse control disorders and rapid eye movement behaviors), dopaminergic uptake (DaT-Scan) and genetic status (APOE, MAPT and SNCA) at baseline and after 5 years. Baseline demographic (age, sex, education years) and clinical values (duration of disease, age of onset) were also included in the model. The algorithm defined (1) the most important variables in predicting MCI, (2) the threshold values to distinguish “converting” vs. “non-converting” subgroups. Results: The algorithm showed that (1) age onset of disease, (2) dopaminergic uptake, (3) age, (4) anxiety, and (5) years of education were the most important factors in predicting MCI over 5 years. Among the factors involved in predicting conversion to MCI, a lower number of years of education associated with lower dopaminergic uptake in the right putamen increased the risk of conversion. Individuals with more years of education are at higher risk of conversion if they have symptoms of depression, anxiety, and lower right striatal dopamine uptake. Other factors that were involved in increasing the risk, were the presence of sleepiness and the presence of rapid eye movement disorders. Interestingly, the genetic factors were of negligible importance and were not considered by the algorithm. Finally, the model showed an accuracy of classification of participants (converters vs. non-converters) of 92.53%. Conclusions: Random forest algorithm shows that (1) depression and anxiety are probably important factors for MCI conversion; (2) years of education influences the conversion; (3) presence of sleepiness and rapid eye movement increases the risk of conversion to MCI. Since the algorithm considers the disease’s age onset, but not the diagnosis of individuals, it would be necessary to generate a model for each group (Healthy on the one hand, Parkinson’s on the other).
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 | 0.002 | 0.003 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.002 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 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".