Currently, steady-state visual evoked potentials (SSVEPs) are applied in a variety of fields. In these applications, spatial filtering is the most commonly used method for decoding SSVEPs. However, existing methods for the SSVEP decoding require improvements in terms of the interaction speed, accuracy, and ease of use. Recent advances in performance have been attributed to the existing hybridisation-based approaches. This article presents and evaluates a novel approach for target identification of SSVEP. The suggested target recognition method significantly enhances the performance of SSVEP decoding by combining two existing methods: extended canonical correlation analysis (eCCA) and the sum of squared correlation (SSCOR). This combined approach is referred to as the extended CCA and sum of squared correlation approach (eCCA-SSCOR). The results of the offline experimental comparison are based on two publicly available datasets, including the San Diego dataset, which comprises 12 frequency targets recorded from 10 individuals, and the Benchmark dataset, comprising 40 frequency targets from 35 subjects. The statistical significance of the improvement was tested using paired samples t-tests and Wilcoxon rank-sum tests. The results indicate that the suggested eCCA-SSCOR approach significantly improves detection accuracy and ITR compared to four existing state-of-the-art methods: canonical correlation analysis (CCA), filter bank CCA (FBCCA), eCCA, and SSCOR. Our method achieved a high average accuracy for all subjects, with 99.5% for the San Diego dataset, and 98.60% for the Benchmark dataset corresponding to data length of 3.5 s and 5 s, respectively. Furthermore, a high ITR of 212.44 bits/min was achieved with a data duration (\(T_w\)) of 0.75 s using the San Diego dataset and was 238.66 bits/min with a \(T_w\) of 1 s for the Benchmark dataset. These offline results demonstrate that the proposed approach successfully combines eCCA and SSCOR to improve SSVEP decoding. Additionally, it is advantageous for applications based on offline processing and can be adapted for online analysis. Online applications can benefit from these results by providing fast interfaces with a large number of commands.

目前，稳态视觉诱发电位（SSVEP）已应用于各个领域。在这些应用中，空间滤波是解码 SSVEP 最常用的方法。然而，现有的 SSVEP 解码方法需要在交互速度、准确性和易用性方面进行改进。最近的性能进步归功于现有的基于杂交的方法。本文提出并评估了一种用于 SSVEP 目标识别的新方法。所提出的目标识别方法通过结合两种现有方法显着增强了 SSVEP 解码的性能：扩展规范相关分析（eCCA）和平方相关和（SSCOR）。这种组合方法称为扩展 CCA 和平方和相关方法 (eCCA-SSCOR)。离线实验比较的结果基于两个公开的数据集，包括圣地亚哥数据集（由 10 个人记录的 12 个频率目标组成）和基准数据集（由 35 名受试者记录的 40 个频率目标组成）。使用配对样本t检验和 Wilcoxon 秩和检验来测试改进的统计显着性。结果表明，与四种现有最先进的方法相比，建议的 eCCA-SSCOR 方法显着提高了检测精度和 ITR：典型相关分析 (CCA)、滤波器组 CCA (FBCCA)、eCCA 和 SSCOR。我们的方法对所有受试者都实现了较高的平均准确率，在圣地亚哥数据集上达到了 99.5%，在数据长度为 3.5 秒和 5 秒的基准数据集上达到了 98.60%。此外，使用圣地亚哥数据集，在数据持续时间 ( \(T_w\) ) 为 0.75 秒时实现了 212.44 位/分钟的高 ITR，而在基准测试中，在\(T_w\)为 1 秒时实现了 238.66 位/分钟。数据集。这些离线结果表明，所提出的方法成功地结合了 eCCA 和 SSCOR 以改进 SSVEP 解码。此外，它对于基于离线处理的应用程序是有利的，并且可以适用于在线分析。在线应用程序可以通过提供具有大量命令的快速接口来从这些结果中受益。