Asst.lect. Haider Raad Hafez
- PhD student at the University of Oldenburg, Germany / Faculty of Medicine / Department of Life Medicine and Medical Information

A research titled: Improving and removing kinetic distortions in the optical brain signals by hybridizing and merging the two wave methods and the enhanced wave correlation.

In the Swiss journal (Sensors ISSN: 1424-8220) affiliated with the MDPI publishing house, with an impact factor of 3.847, and classified within the Scopus index Q1 database, as part of the requirements for an international doctoral degree within Scopus repositories.

Where did the research take place
Near infrared spectroscopy (fNIRS) is a non-invasive optical neuroimaging technique that allows participants to move relatively freely. However, head movements often cause motions of the electrodes relative to the head, which introduce motion distortions in the measured signal. Here, we propose an improved algorithmic approach to correct those kinematic distortions that combines wave-based and correlation-signal optimization (WCBSI). We compare the accuracy of MA correction of its kinematic distortions with multiple established correction approaches (linear interpolation, spline-Savitzky-Golay filter, principal component analysis, target principal component analysis, locally weighted strong gradient smoothing filter, wavelet filter, and correlation-based signal optimization). ). Therefore, we measured brain activity in 20 participants performing a task of moving their hands on a table and moving their heads simultaneously to produce motor abnormalities of different levels of severity. In order to obtain the 'true cue' of brain activation, we added a condition in which the hand-moving task was performed on a table only. We compared the distortion correction performance between the algorithms on four predefined metrics (R, RMSE, MAPE, and AUC) and ranked the performance. The proposed WCBSI algorithm was the only one that outperformed the mean (p < 0.001), and had the highest probability of being the best-ranked algorithm (probability 78.8%). Together, our results indicate that among all the algorithms tested, the proposed WCBSI approach performed consistently well across all scales.