時間預期對工作記憶反應的調節：行為與電生理證據

Modulation of Working Memory Responses by Temporal Expectations: Behavioural and Electrophysiological Evidence

許多研究顯示時間預期可以幫助個體預測事件的發生時間，影響行為表現及神經動態反應。儘管此效應在知覺中已被廣泛發現，但對工作記憶測試時間的預期是否同樣能調節工作記憶的反應仍不清楚。本研究透過三個行為實驗與一個腦電圖實驗探討此問題，並在延遲反應工作記憶任務中操弄延遲時間變異性。不同於以往研究在單次嘗試使用提示線索，我們透過操弄樣本記憶與測試探針刺激之間的延遲時長，設置了小變異性與大變異性條件，在脈絡層級上誘發不同的預期狀態。在實驗一中，我們發現當延遲時長變異性較小時，相較於變異性較大，反應時間顯著變快，顯示時間可預測性提升了反應效率。而此效益在特定固定延遲時長的情況下消失（實驗二A與二B）。在實驗三中，我們在腦電圖實驗重現實驗一的發現，並觀察到在小變異性條件下，測試探針出現時反應手對側的α波功率下降，反映了更有效的反應選擇過程。本結果顯示，當測試探針的時間較為可預測時，時間預期能動態地透過增強延遲反應的選擇與準備來提升表現。這些發現不僅深化了我們對時間預期性優化認知表現的理解，亦顯示出時間可預測性在工作記憶中的促進效益可由α波的調控作為指標。本研究揭示了時間預期在引導工作記憶反應中的主動調節作用。

Many studies have shown that temporal expectations influence behavioural performance and neural dynamics by enabling individuals to anticipate the timing of upcoming events. While such effects have been consistently demonstrated in perception, it remains unclear whether temporal expectations regarding the onset of a memory test can similarly modulate working memory (WM) responses. The current study addressed this question across three behavioural experiments and one electroencephalography (EEG) experiment, manipulating delay duration variability in a delayed-response WM task. Unlike previous studies that used trial-level cues, we manipulated delay durations between the sample and test probe to establish conditions with small or large variability at the contextual level, thereby leading to distinct anticipatory states. In Experiment 1, we demonstrated that WM responses were faster when delay durations varied within a narrow range (small variability) compared to a broader range (large variability), suggesting that temporal predictability enhances response efficiency. This benefit diminished when delay durations were fixed at specific intervals (experiments 2A and 2B), indicating that temporal variability is a critical factor in modulating WM performance. In experiment 3, we replicated the findings of Experiment 1 using EEG and observed that small variability trials elicited a temporally specific reduction in alpha power contralateral to the response hand, indicative of efficient response selection. Our findings suggest that temporal expectations dynamically modulate WM processes by enhancing the selection and preparation of delayed responses when the timing of the test probe is more predictable. These results advance our understanding of how temporal cues optimise cognitive performance and provide evidence that the beneficial effects of temporal predictability on WM can be indexed by alpha oscillations. Together, this work highlights the proactive role of temporal expectation in guiding WM responses.