基于双耳分听范式的听觉脑机接口研究

An auditory brain-computer interface based on dichotic listening paradigm

为了探寻高效的听觉脑机接口实验范式及有效的数据处理方法,本文设计了双耳分听范式.在被试左右耳朵同时播放两个不同的声音刺激序列,每个声音序列中都包含靶刺激和非靶刺激,被试通过关注靶刺激来关注相应方向的声音序列.靶刺激诱发了事件相关电位P300,通过判断P300出现时间即可确定被试关注声音序列的方向.通过一维离散小波变换对脑电数据进行单次样本特征提取,选取低频的特征向量,并用支持向量机进行目标与非目标的识别.最终目标识别正确率达到80%以上,能够与视觉脑机接口相比拟,可以用于二分类的脑机接口系统.

This study proposed a dichotic listening paradigm in order to explore an effective auditory Brain-computer InAn Auditory Brain-computer Interface Based on Dichotic Listening Paradigm Abstract The auditory brain-computer interface system brings good news for patients with hypopsia or locked-in syndrome who cannot control eye movement. In this paper, it designs a Dichotic Listening experimental paradigm by playing different sound stimulation sequences to both ears at the same time. Then participants make a choice through paying attention to sound sequence in the left ear or right ear. Due to having made an agreement on the corresponding option of left ear or right ear in advance, the participants can focus on the source of the sound earlier in the experimental process for a better effect. By knowing when event-related potentials P300 appeared, we judge the results of what participants choose to focus on. For data analysis, we use wavelet transform method for feature extraction and support vector machine(SVM) to identify the target and non-target. The accuracy of the ultimate target identification can reach more than 80%, which can be used for binary-class of brain-computer interface system.