Modified Approach for Solving Hemodynamic Model by fMRI Images

The hemodynamic model is a crucial method for effectively monitoring and potentially predicting brain activity. Researching the solution of the hemodynamic model is a feasible task. Three popular techniques in the field of biomedical engineering are utilised to solve the hemodynamic model parameters, hidden states, and estimate blood oxygen level dependent (BOLD) signal from fMRI pictures. The dynamic expectation maximisation (DEM) is a commonly used technique for inverting a hemodynamic model. Both the square root cubature Kalman smoothing (SCKS) and the recently introduced confounds SCKS (CSCKS) are derived from the cubature Kalman filtering (CKF) approach. The standard algorithms provide the necessary measurement, but their accuracy is lower compared to the suggested new CSCKS technique, which has improved algorithms. The new CSCKS algorithm provides hidden state estimates that are 89% accurate, while the old CSCKS, SCKS, and DEM methods provide estimates that are 82%, 53%, and 36% accurate, respectively. Under identical parameter conditions, the new CSCKS algorithm demonstrates an accuracy that is 5%, 1.5%, and 16.07% higher than the old CSCKS method in estimating biophysical parameters, signal to interference ratio (SIR) of the BOLD response, and vascular signal of blood flow, respectively.