BACKGROUND: Major depressive disorder (MDD) is characterized by pervasive cognitive-emotional biases, yet the spatiotemporal dynamics of prefrontal involvement during emotional association remain poorly understood. This study aimed to delineate temporal deviations and spatial reweighting of prefrontal activation in MDD and evaluate their relationship to behavioral bias and clinical symptom burden. METHODS: Sixty-nine participants (47 MDD, 22 healthy controls) completed the Free Association Semantic Task (FAST) during functional near-infrared spectroscopy (fNIRS) recording. Participants generated 10 associations in response to neutral, positive, or negative cue words. Emotional valence ratings were derived for each association, and group differences were assessed using independent-samples t-tests. Prefrontal hemodynamic responses were preprocessed and compared across six regions of interest using Bonferroni-corrected analyses. Spearman correlations examined brain-behavior relationships between valence trajectories, prefrontal activation, and clinical measures. RESULTS: MDD patients exhibited consistent negative valence drift under neutral and positive cues, but not under negative cues. fNIRS revealed distinct temporal deviations, characterized by insufficient early prefrontal activation followed by delayed compensatory recruitment, and spatial deviations, with exaggerated reliance on medial prefrontal cortex (mPFC) and attenuated dorsolateral prefrontal cortex (dlPFC) modulation. These spatiotemporal patterns persisted even in the absence of behavioral differences under negative cues. Brain-behavior analyses showed that stronger late negative associations correlated with higher depressive severity and insomnia, whereas increased mPFC and temporal activation reflected compensatory attempts at regulation. CONCLUSIONS: MDD is marked by disrupted spatiotemporal prefrontal signatures, including delayed and prolonged activation and spatial imbalance favoring mPFC over dlPFC. These deviations provide mechanistic insight into depressive cognitive bias and nominate temporal-spatial prefrontal dynamics as ecological biomarkers with potential utility for precision psychiatry. CLINICAL TRIAL NUMBER: Not applicable.