Kuo C.-H.Kuo Y.-C.Chou H.-C.Lin Y.-T.CHUNG-HSIEN KUO2022-05-242022-05-242017https://www.scopus.com/inward/record.uri?eid=2-s2.0-85016807515&doi=10.1007%2fs40815-016-0205-x&partnerID=40&md5=6553088c1d2cd600220765568423aacehttps://scholars.lib.ntu.edu.tw/handle/123456789/611562P300 is a brain–computer interface (BCI) modality which reflects brains’ processes in stimulus events. Visual stimuli are usually used to elicit event-related P300 components. However, depending on different subjects’ conditions and their current cerebral loads, P300 components occur at posterior to stimuli from 250 to 600?ms roughly. These subjects’ dependent variations affect the performance of BCI. Thus, an estimation model that estimated an appropriate interval for P300 feature extraction is discussed in this paper. An interval type-2 fuzzy logic system trained by artificial bee-colony algorithm was used to find the latency of elicited P300 with a certain range by means of steady-state visually evoked potential. A support vector machine classifier was adopted to classify extracted epochs into target and non-target stimuli. Seven subjects were involved in experiments. Results showed that the performance of information transfer rate was improved by 1.28?% on average if the proposed latency-estimation approach was introduced. ? 2016, Taiwan Fuzzy Systems Association and Springer-Verlag Berlin Heidelberg.Bioelectric potentialsComputer circuitsEvolutionary algorithmsFeature extractionFuzzy logicInterface statesOptimizationArtificial bee colony algorithmsInformation transfer rateInterval type-2 fuzzy logic systemsLatency estimationP300Steady state visually evoked potentialsSupport vector machine classifiersTarget and non targetsBrain computer interfacejournal article10.1007/s40815-016-0205-x2-s2.0-85016807515