BCIwiki (bciwiki.com) — A researcher at Imam Ja'afar Al-Sadiq University in Baghdad has proposed a new multi-criteria decision framework for identifying the 'golden subject' in motor imagery brain-computer interface (MI-BCI) applications, as reported in the Mesopotamian Journal of Computer Science on September 22, 2026. In cross-subject transfer learning, the golden subject is the individual whose EEG patterns generalize best to other users. An earlier iteration of the same research line, the HAFL study, used the fuzzy decision by opinion score method (FDOSM) and selected Subject 5 (S5) as the top candidate. The new study separates criteria weighting from compromise ranking by building a unified 9 x 18 Golden Subject Decision Matrix (UGSDM) that evaluates nine subjects across 18 criteria spanning both training and evaluation phases, including classification accuracy, recall, F1-score, log loss, and specificity.
FWZIC was applied to compute criteria weights, and VIKOR was used to generate a compromise ranking. S1 emerged as the best compromise subject with Q = 0.0000, followed by S3 (Q = 0.1722) and S5 (Q = 0.3373) — a reversal from the earlier FDOSM result that had placed S5 first. Sensitivity analysis across four weighting scenarios confirmed the result: S1 ranked first in three scenarios and second only in the training-dominant scenario. The author argues that the ranking shift demonstrates how combining training and evaluation criteria in a single matrix, under different weighting logic, can surface different subject priorities. The study is limited to nine subjects and has not been validated on larger public datasets.