# Predictive model of neurocognitive functioning after acute coronary syndrome. A machine learning approach

## TL;DR

This study uses machine learning to predict neurocognitive functioning in patients with acute coronary syndrome, identifying key risk factors like depression and cholesterol.

## Contribution

A novel machine learning model is developed to predict neurocognitive outcomes in acute coronary syndrome patients.

## Key findings

- The random forest model achieved high accuracy with an r-squared of 0.978.
- Key predictors included HDL, depression, glucose, and BMI, highlighting cardiovascular and mental health factors.
- The model supports early identification of patients needing tailored interventions.

## Abstract

The interplay between coronary disease and neurocognitive dysfunction remains unclear with several underlying factors likely contributing to this complex relationship. This study develops a predictive model using a machine learning approach to determine a predictive model of neurocognitive functioning in patients with acute coronary syndrome (ACS).

Sixty-three patients, enrolled in the phase III cardiac rehabilitation program, underwent a neurocognitive assessment. To predict neurocognitive functioning a cross validated random forest model was used (RF_cv) due to its robustness to non-linear relationships and overfitting, and its successful application in prior disease prediction studies.

The RF_cv model showed an r-squared of 0.978, an RMSE of 0.6309 and a MAE value of 0.479. The top-ten predictors in the model were: HDL, Depression, Glucose, Glycated Hemoglobin, B-Type Natriuretic Peptide, BMI (Kg/m2), Waist-to-Hip Ratio, Cholesterol, Anxiety and Age.

The variance in neurocognitive functioning is explained by a combination of biochemical indicators and body composition, reflecting classical cardiovascular risk factors and depression. The obtained RF-cv predictive model supports early identification of patients for tailored interventions.