Early stress detection using wearable sensors requires understanding which modality responds first. Most studies compare stress versus baseline as static epochs, discarding temporal information. We investigated the temporal dynamics of ECG, EDA, and PPG stress biomarkers, including the anticipatory phase. We analyzed the WESAD dataset (15 subjects) recorded during the Trier Social Stress Test. Six features were selected: heart rate (HR) and the root mean square of successive RR-interval differences (RMSSD) from ECG; skin conductance level (SCL) and skin conductance response (SCR) rate from EDA; and normalized pulse area and pulse skewness from PPG. In sliding 30-s windows we computed paired Cohen's d against a 5-min resting baseline, including the preceding transitions. ECG responded earliest: heart rate rose already during the anticipatory transition before the formal stress onset (d = 1.6 during stress), an elevation not explained by movement (wrist accelerometry). EDA responded more slowly. PPG showed strong effects (pulse area d = −2.3), though with higher variability. Meditation produced a mirror-image pattern; amusement and baseline showed minimal effects, confirming stress specificity. ECG captures the stress response earlier than EDA, emerging already during the anticipatory phase, and appears most suitable for early wearable stress detection.