Predictive coding framework posits that our brain continuously monitors changes in the environment and updates its predictive models, minimizing prediction errors to efficiently adapt to environmental demands. However, the underlying neurophysiological mechanisms of these predictive phenomena remain unclear. The present study aimed to explore the systemic neurophysiological correlates of predictive coding processes during passive and active auditory processing. Electroencephalography (EEG), functional near-infrared spectroscopy (fNIRS), and autonomic nervous system (ANS) measures were analyzed using an auditory pattern-based novelty oddball paradigm. A sample of 32 healthy subjects was recruited. The results showed shared slow evoked potentials between passive and active conditions that could be interpreted as automatic predictive processes of anticipation and updating, independent of conscious attentional effort. A dissociated topography of the cortical hemodynamic activity and distinctive evoked potentials upon auditory pattern violation were also found between both conditions, whereas only conscious perception leading to imperative responses was accompanied by phasic ANS responses. These results suggest a systemic-level hierarchical reallocation of predictive coding neural resources as a function of contextual demands in the face of sensory stimulation. Principal component analysis permitted to associate the variability of some of the recorded signals.

中文翻译：

---

听觉预测编码的系统神经生理信号

预测编码框架假设我们的大脑不断监测环境的变化并更新其预测模型，最大限度地减少预测错误，以有效地适应环境需求。然而，这些预测现象的潜在神经生理学机制仍不清楚。本研究旨在探索被动和主动听觉处理过程中预测编码过程的系统神经生理学相关性。使用基于听觉模式的新奇古怪范例对脑电图 (EEG)、功能性近红外光谱 (fNIRS) 和自主神经系统 (ANS) 测量进行分析。招募了 32 名健康受试者作为样本。结果显示，被动和主动条件之间共享慢诱发电位，可以解释为预期和更新的自动预测过程，独立于有意识的注意力努力。在两种情况下也发现了皮质血流动力学活动的分离地形和听觉模式违反时独特的诱发电位，而只有导致命令性反应的有意识知觉才会伴随阶段性 ANS 反应。这些结果表明，在面对感觉刺激时，预测编码神经资源作为上下文需求的函数进行了系统级的分层重新分配。主成分分析允许将一些记录信号的变异性关联起来。