Back

A machine learning approach for MRI-based classification of individuals with mild cognitive impairment

Tibon, R.

2021-05-28 neuroscience
10.1101/2021.05.27.445930 bioRxiv
Show abstract

Diagnosis criteria for Mild Cognitive Impairment (MCI) rely on cognitive symptoms and do not require specific biomarker evidence of pathology. Nevertheless, patients may still have subtle brain changes that are identifiable with neuroimaging. To classify MCI patients vs. healthy age-matched controls, two algorithms--Multilayer Perceptron (MLP) and Support Vector Machine (SVM)--were applied to structural MRI data from the OASIS-1 cohort (https://www.oasis-brains.org/). Despite their comparable performance in some measures, the MLP algorithm was preferable due to superior recall of MCI cases.

Matching journals

The top 8 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.