A Framework for Precise ERP Detection by Compensating for Perceptual Timing Misalignment

Accurate detection of event-related potentials (ERPs) in tactile experiments is challenging because the perceptual onset of tactile stimuli varies substantially across trials and individuals. This perceptual timing variability disrupts temporal alignment during conventional averaging, resulting in smeared ERP waveforms and unreliable component detection. To address this problem, we propose a two-stage ERP extraction framework consisting of a correlation-based trigger alignment method and a Modified Matching Method (MMM). The correlation-based trigger alignment method effectively identifies optimal timing shifts by maximizing correlation coefficients between EEG signals, thereby mitigating the trial-to-trial jitter inherent in tactile perception and ensuring that ERP components are precisely synchronized before averaging. The second stage, MMM, further enhances the signal-to-noise ratio by incorporating pre-averaging and data augmentation. The proposed approach was evaluated through auditory and tactile oddball paradigms. In the auditory experiment, the method improved ERP clarity for small pre-averaging sizes (N=2–3), outperforming the conventional averaging method. In the tactile experiment, where perceptual timing variability was larger, the framework demonstrated even greater effectiveness, successfully recovering ERP components that were not detectable using traditional averaging. A structured scoring system confirmed that the proposed framework compensates for perceptual timing misalignment through precise trigger matching and enhances ERP extraction, particularly in tactile modalities. This study demonstrates the feasibility of using EEG to evaluate tactile perception and provides a foundation for future neural-based assessments of tactile processing.

触觉实验中事件相关电位（Event-related Potentials, ERPs）的精准检测颇具挑战，原因在于触觉刺激的知觉起始时刻在不同试次与个体间均存在显著差异。这种知觉计时变异性会干扰传统平均法中的时间对齐过程，进而导致ERP波形模糊、成分检测结果不可靠。为解决该问题，本文提出一种两阶段ERP提取框架，包含基于相关性的触发对齐方法与改进匹配法（Modified Matching Method, MMM）。该基于相关性的触发对齐方法通过最大化脑电信号（Electroencephalogram, EEG）间的相关系数，有效识别最优时移，从而缓解触觉知觉固有的试次间抖动问题，确保平均操作前ERP成分实现精准同步。第二阶段的改进匹配法可通过引入预平均与数据增强操作，进一步提升信噪比。本文通过听觉与触觉异常刺激范式（oddball paradigm）对所提方法开展了性能评估。在听觉实验中，当预平均样本量较小时（N=2~3），该方法可提升ERP波形清晰度，性能优于传统平均法。在知觉计时变异性更大的触觉实验中，该框架展现出更优异的性能，成功还原了传统平均法无法检测到的ERP成分。通过结构化评分系统验证可知，所提框架可通过精准的触发匹配补偿知觉计时错位问题，进而提升ERP提取效果，尤其在触觉模态下效果显著。本研究证实了采用脑电评估触觉知觉的可行性，并为未来基于神经信号的触觉加工评估奠定了基础。