Loading...
2 results
Search Results
Now showing 1 - 2 of 2
- Impact of age, VR, immersion, and spatial resolution on classifier performance for a MI-based BCIPublication . Blanco-Mora, D. A.; Aldridge, A.; Jorge, C.; Vourvopoulos, A.; Figueiredo, P.; Bermúdez i Badia, S.There are many factors outlined in the signal processing pipeline that impact brain–computer interface (BCI) performance, but some methodological factors do not depend on signal processing. Nevertheless, there is a lack of research assessing the effect of such factors. Here, we investigate the impact of VR, immersiveness, age, and spatial resolution on the classifier performance of a Motor Imagery (MI) electroencephalography (EEG)-based BCI in naïve participants. We found significantly better performance for VR compared to non-VR (15 electrodes: VR 77.48 ± 6.09%, non-VR 73.5 ± 5.89%, p = 0.0096; 12 electrodes: VR 73.26 ± 5.2%, non-VR 70.87 ± 4.96%, p = 0.0129; 7 electrodes: VR 66.74 ± 5.92%, non-VR 63.09 ± 8.16%, p = 0.0362) and better performance for higher electrode quantity, but no significant differences were found between immersive and non immersive VR. Finally, there was not a statistically significant correlation found between age and classifier performance, but there was a direct relation found between spatial resolution (electrode quantity) and classifier performance (r = 1, p = 0.0129, VR; r = 0.99, p = 0.0859, non-VR).
- Finding the optimal time window for increased classification accuracy during motor imageryPublication . Blanco-Mora, D. A.; Aldridge, A.; Jorge, C.; Vourvopoulos, A.; Figueiredo, P.; Bermúdez i Badia, S.Motor imagery classification using electroencephalography is based on feature extraction over a length of time, and different configurations of settings can alter the performance of a classifier. Nevertheless, there is a lack of standardized settings for motor imagery classification. This work analyzes the effect of age on motor imagery training performance for two common spatial pattern-based classifier pipelines and various configurations of timing parameters, such as epochs, windows, and offsets. Results showed significant (p ≤ 0.01) inverse correlations between performance and feature quantity, as well as between performance and epoch/window ratio.