Sensory processing sensitivity (SPS) is a biological trait associated with heightened sensitivity and responsivity to the environment. One important question is how those with the trait perceive their environments, thus giving rise to differential responses and outcomes. In this study, we used an artificial intelligence (AI) model—SPS Vision Net—to investigate perceptual differences associated with SPS and to begin to predict sensitivity levels based on a visual perception task. 190 participants (M age = 22.91; 102 (53%) females), completed an online experiment where they rated 100 images from the Open Affective Standardized Image Set (OASIS) on arousal, valence, and visual saliency. They also completed the Highly Sensitive Person (HSP) Scale measure of SPS. Results showed that SPS was positively associated with arousal in response to negative (vs. positive and neutral images), and, namely, sad (vs. happy, neutral, or fear) images. Also, SPS was negatively associated with positive ratings of negative images, specifically those showing frightening images. SPS was unrelated to response times and the number of salient selection blocks made. However, the AI model showed high accuracy (83.31%) in predicting SPS levels (R² = 0.77). Consistent with theory and research, this study showed that SPS is associated with higher arousal and lower positive ratings in response to the OASIS image rating task. Novel findings showed that a new, accurate AI-backed SPS measurement system, based on a visual selection, was predictive of HSP scores with high accuracy. Finally, the AI model indicates that visual perception differs as a function of SPS.

感觉处理敏感性（SPS）是一种与对环境的敏感性和反应性增强相关的生物特征。一个重要的问题是具有这一特征的人如何感知他们的环境，从而产生不同的反应和结果。在这项研究中，我们使用人工智能 (AI) 模型——SPS Vision Net——来研究与 SPS 相关的感知差异，并开始根据视觉感知任务预测敏感度水平。190 名参与者（M年龄 = 22.91；102 名（53%）女性）完成了一项在线实验，他们对开放情感标准化图像集 (OASIS) 中的 100 张图像的唤醒度、效价和视觉显着性进行了评分。他们还完成了 SPS 的高度敏感人群 (HSP) 量表测量。结果表明，SPS 与消极（相对于积极和中性图像）以及悲伤（相对于快乐、中性或恐惧）图像的反应呈正相关。此外，SPS 与负面图像的正面评级呈负相关，特别是那些显示可怕图像的正面评级。SPS 与响应时间和显着选择块的数量无关。然而，AI 模型在预测 SPS 水平方面表现出较高的准确度 (83.31%) ( R ²  = 0.77)。与理论和研究一致，本研究表明，SPS 与 OASIS 图像评级任务的较高唤醒度和较低积极评级相关。新的研究结果表明，基于视觉选择的全新、准确的人工智能支持的 SPS 测量系统可以高精度预测 HSP 分数。最后，AI 模型表明视觉感知随着 SPS 的变化而变化。