Artificial intelligence for diagnosis and prognosis in neuroimaging for dementia; a systematic review
Bibliographic record
Abstract
Abstract Introduction Recent developments in artificial intelligence (AI) and neuroimaging offer new opportunities for improving diagnosis and prognosis of dementia. To synthesise the available literature, we performed a systematic review. Methods We systematically reviewed primary research publications up to January 2021, using AI for neuroimaging to predict diagnosis and/or prognosis in cognitive neurodegenerative diseases. After initial screening, data from each study was extracted, including: demographic information, AI methods, neuroimaging features, and results. Results We found 2709 reports, with 252 eligible papers remaining following screening. Most studies relied on the Alzheimer’s Disease Neuroimaging Initiative (ADNI) dataset (n=178) with no other individual dataset used more than 5 times. Algorithmic classifiers, such as support vector machine (SVM), were the most commonly used AI method (47%) followed by discriminative (32%) and generative (11%) classifiers. Structural MRI was used in 71% of studies with a wide range of accuracies for the diagnosis of neurodegenerative diseases and predicting prognosis. Lower accuracy was found in studies using a multi-class classifier or an external cohort as the validation group. There was improvement in accuracy when neuroimaging modalities were combined, e.g. PET and structural MRI. Only 17 papers studied non-Alzheimer’s disease dementias. Conclusion The use of AI with neuroimaging for diagnosis and prognosis in dementia is a rapidly emerging field. We make a number of recommendations addressing the definition of key clinical questions, heterogeneity of AI methods, and the availability of appropriate and representative data. We anticipate that addressing these issues will enable the field to move towards meaningful clinical translation.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.003 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.003 | 0.001 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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 teacher head, 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".