Department of Special Education and Communication Disorders, University of Nebraska-Lincoln, Lincoln, NE, USA.
Department of Electrical and Computer Engineering, University of Nebraska-Lincoln, Lincoln, NE, USA.
Neuropsychologia. 2022 Aug 13;173:108289. doi: 10.1016/j.neuropsychologia.2022.108289. Epub 2022 Jun 9.
Brain-computer interfaces for augmentative and alternative communication (BCI-AAC) may help overcome physical barriers to AAC access. Traditionally, visually based P300-BCI-AAC displays utilize a symmetrical grid layout. Contextual scene displays are composed of context-rich images (e.g., photographs) and may support AAC success. However, contextual scene displays contrast starkly with the standard P300-grid approach. Understanding the neurological processes from which BCI-AAC devices function is crucial to human-centered computing for BCI-AAC. Therefore, the aim of this multidisciplinary investigation is to provide an initial exploration of contextual scene use for BCI-AAC.
Participants completed three experimental conditions to evaluate the effects of item arrangement asymmetry and context on P300-based BCI-AAC signals and offline BCI-AAC accuracy, including 1) the full contextual scene condition, 2) asymmetrical item arraignment without context condition and 3) the grid condition. Following each condition, participants completed task-evaluation ratings (e.g., engagement). Offline BCI-AAC accuracy for each condition was evaluated using cross-validation.
Display asymmetry significantly decreased P300 latency in the centro-parietal cluster. P300 amplitudes in the frontal cluster were decreased, though nonsignificantly. Display context significantly increased N170 amplitudes in the occipital cluster, and N400 amplitudes in the centro-parietal and occipital clusters. Scenes were rated as more visually appealing and engaging, and offline BCI-AAC performance for the scene condition was not statistically different from the grid standard.
Findings support the feasibility of incorporating scene-based displays for P300-BCI-AAC development to help provide communication for individuals with minimal or emerging language and literacy skills.
未加标签:用于辅助和替代交流的脑机接口(BCI-AAC)可以帮助克服交流辅助设备获取的物理障碍。传统上,基于视觉的 P300-BCI-AAC 显示器采用对称网格布局。语境场景显示由富有语境的图像(例如照片)组成,可以支持 AAC 的成功。然而,语境场景显示与标准的 P300 网格方法形成鲜明对比。了解脑机接口设备运行的神经过程对于以人类为中心的脑机接口计算至关重要。因此,这项多学科研究的目的是初步探索语境场景在 BCI-AAC 中的应用。
方法:参与者完成了三个实验条件,以评估项目排列不对称和上下文对基于 P300 的 BCI-AAC 信号和离线 BCI-AAC 准确性的影响,包括 1)完整语境场景条件,2)无上下文的不对称项目排列条件和 3)网格条件。在每个条件之后,参与者完成了任务评估评分(例如参与度)。使用交叉验证评估每个条件的离线 BCI-AAC 准确性。
结果:显示不对称显著降低了中央顶叶簇中的 P300 潜伏期。额区的 P300 振幅虽略有下降但无统计学意义。显示上下文显著增加了枕区的 N170 振幅,以及中央顶叶和枕区的 N400 振幅。场景被评为更具视觉吸引力和吸引力,并且场景条件下的离线 BCI-AAC 性能在统计学上与网格标准没有差异。
结论:研究结果支持将基于场景的显示纳入 P300-BCI-AAC 开发,以帮助为语言和读写能力有限或正在发展的个人提供交流。