目的：本研究旨在探討不同踩踏頻率 (50、70、90、110 rpm) 下，騎乘公路自行車之上肢肌肉活化程度 (尺側伸腕肌、橈側屈腕肌、第一至第四背側骨間肌) 與下肢膝關節偏移量及穩定度之差異，並分析其之間的相關性。方法：招募30位男性受試者，進行功率150 W四種頻率的騎乘。以無線表面肌電儀及透過慣性感測器收取實驗數據。分別以重複量數單因子變異數分析及Bonferroni法事後比較分析不同踏頻之差異，另以斯皮爾曼等級相關進行之間的相關性統計與二次迴歸曲線估計，顯著水準設為α = .05。結果：踩踏偏移量在110 rpm顯著較高 (左：50, 70, 90 < 110 rpm；右：50, 70 < 90 < 110 rpm)，而穩定度則是50 rpm顯著較高 (左：50 > 110 rpm；右：50 > 70, 90, 110 rpm; 70 > 110 rpm)。上肢僅伸腕肌在110 rpm時肌肉活化顯著較高 (110 > 50, 70, 90 rpm)。而50 rpm中有較明顯的上下肢相關性，DI-1與DI-3與雙腳偏移量皆有中等正相關，伸腕肌則是與雙腳穩定度接近於中等正相關。結論：本研究發現高踏頻會因神經肌肉控制能力下降而導致膝關節穩定度下降及前臂肌肉活化上升；低踏頻則增加了上下肢的相關性，反映出上肢潛在的補償機制。不僅補充了上下肢互動影響的實證資料，並提供了騎乘機制與踏頻選擇的依據，對於訓練規劃與姿勢調整具重要參考價值。

Purpose: This study investigated differences in upper limb muscle activation—extensor carpi ulnaris (ECU), flexor carpi radialis (FCR), and first to fourth dorsal interossei (DI-1 to DI-4)—and knee joint mediolateral deviation and stability across four pedaling cadences (50, 70, 90, 110 rpm) during road cycling. It also assessed correlations between these variables. Methods: Thirty male participants completed cycling trials at 150 W under four pedaling cadences. Data were collected using wireless surface EMG and inertial measurement units. Repeated-measures ANOVA with Bonferroni post hoc tests assessed cadence differences. Spearman’s rank correlation and quadratic regression were used to analyze variable relationships. Significance was set at α = .05. Results: pedaling deviation was significantly higher at 110 rpm (Left: 50, 70, 90 < 110 rpm; Right: 50, 70 < 90 < 110 rpm), while stability was greater at 50 rpm (Left: 50 > 110 rpm; Right: 50 > 70, 90, 110 rpm; 70 > 110 rpm). Among upper limb muscles, only the ECU showed significantly higher activation at 110 rpm (110 > 50, 70, 90 rpm). At 50 rpm, upper-limb and lower-limb correlations were more pronounced: DI-1 and DI-3 were moderately positively correlated with bilateral pedaling deviation, and ECU showed a near-moderate positive correlation with bilateral stability. Conclusion: High pedaling cadence led to reduced knee stability and increased forearm muscle activation, likely due to limited neuromuscular control. Low cadence showed stronger upper–lower limb correlations, indicating potential upper-limb compensation. These results provide empirical support for limb interaction and inform cadence selection, training design, and posture adjustment.