Evidence record · UDINUS · SDGs
- Unit
- Rektorat UDINUS
- Period
- 2024
- Goals
- Record
- 374a6e43-dc77-4ac2-a83d-760a510e070b
- Published
- 2026-10-30
Universitas Dian Nuswantoro conducted a study addressing the assessment of stroke severity through EEG signal analysis. The activity involved Sulistyono, Marcelinus Yosep Teguh; Pane, Evi Septiana; Yuniarno, Eko Mulyanto; and Purnono, Mauridhi Hery as participants in the research activity. The study was motivated by the role of stroke-severity assessment in determining appropriate medical interventions and supporting the monitoring, evaluation, and rehabilitation process for stroke patients. It identified that neurologists relied solely on visualization during examination sessions, through which significant information could sometimes be overlooked.
The activity examined the use of EEG analysis to supplement visualization for stroke detection and severity assessment. The source describes EEG signals as non-stationary, non-linear, multi-band, and multi-channel. It also identifies difficulties arising from separate processing and feature extraction in each domain, which produces a high-dimensional space. To address these challenges, the study proposed the Hybrid Significant Stroke Feature (HSSF) method and a multi-domain hybrid significant stroke feature set.
The HSSF method used sampling of alpha low and alpha high sub-bands within the alpha band, as well as beta low and beta high sub-bands within the beta band. The approach utilized ten characteristics across the time domain, frequency domain, and signal decomposition domain, together with motion stimuli involving the shoulder, elbow, and grab. Its signal-processing and feature-analysis strategy comprised data collection, data preparation, feature extraction, and selection of salient characteristics for categorization. Dimensionality reduction prioritized low-dimensional domains. Accuracy enhancement involved normalcy testing and significant feature testing to determine the most significant mean value.
The experimental results reported that HSSF classification based on type, domain, feature, stimulus, and subband achieved 98% accuracy in identifying stroke severity through SVM classification using PSD, Hjorth Parameter Mobility, and Hjorth Parameter Complexity features with the high beta sub-band. The results indicate that the method identified optimal parameters within a restricted dimensional space for precise stroke-severity assessment.