Toward FRP-Based Brain-Machine Interfaces - Single-Trial Classification of Fixation-Related Potentials

The co-registration of eye tracking and Electroencephalography (EEG) provides a holistic measure of ongoing cognitive processes. Recently, \emph{Fixation-related potentials} (FRP) have been introduced to quantify the neural activity in these bi-modal recordings. FRPs are time-locked to fixation onsets, just like event-related potentials (ERP) are locked to stimulus onsets. The merit of the FRPs is that they can be analyzed in free viewing conditions and classified on a \emph{single-trial} level. However, existing research has investigated FRPs only with very restricted and highly unnatural stimuli in simple search tasks. We present a study where we relieved many of these restrictions while retaining some control by using a gaze-contingent search task with complex and everyday objects. We obtained high classification accuracies even under the more complex and less constrained task. Furthermore, we show that our classification approach generalizes not only to different test sets from the same participant, but also across participants. These results promise to open an avenue for exploiting FRPs in EEG-based brain-machine interfaces and thus providing a novel means for intuitive human-machine interaction, for example, by providing cues to the machine which objects in a scene are relevant for a user without the need to communicate that explicitly.

眼动追踪与脑电图（Electroencephalography，EEG）的联合记录为正在进行的认知过程提供了整体测量手段。近期，注视相关电位（Fixation-related potentials，FRP）被引入以量化这些双模态记录中的神经活动。FRP与注视起始时刻时间锁定，正如事件相关电位（event-related potentials，ERP）与刺激起始时刻锁定一样。FRP的优势在于其可在自由观看条件下分析，并能在单试次（single-trial）水平上分类。然而，现有研究仅在简单搜索任务中使用高度受限且极不自然的刺激对FRP进行了研究。我们开展了一项研究，通过采用注视依赖的（gaze-contingent）搜索任务并使用复杂日常物体，在保留一定控制的同时解除了诸多此类限制。即使在更复杂、约束更少的任务条件下，我们仍获得了较高的分类准确率。此外，我们发现我们的分类方法不仅能推广到同一被试的不同测试集，还能跨被试推广。这些结果有望为在基于EEG的脑机接口（brain-machine interfaces）中利用FRP开辟新途径，从而为直观的人机交互（human-machine interaction）提供新颖手段——例如，无需用户明确告知，即可向机器提示场景中哪些物体与用户相关。