This study demonstrates that whole-brain functional connectivity patterns can be used to decode subject-driven cognitive states. By analyzing 90 functionally defined regions of interest (ROIs) across 14 large-scale resting-state brain networks, researchers developed a classifier that accurately identifies four cognitive states: undirected rest, episodic memory retrieval, serial subtraction, and silent singing. The classifier achieved 84% accuracy in a leave-one-out cross-validation and 85% accuracy in an independent cohort. It maintained high accuracy even with short imaging runs as brief as 30–60 seconds. Functional ROIs outperformed structural ROIs in classifying cognitive states across all tested durations. The study also tested the classifier's ability to reject a novel cognitive state, spatial navigation, and found it to be highly specific. The results suggest that whole-brain functional connectivity analysis can provide insights into naturalistic, continuous cognitive processing, offering a promising approach for decoding a wide range of subject-driven cognitive states from brief imaging data. The study highlights the potential of functional connectivity in understanding cognitive states without relying on explicit temporal markers, which is a significant advancement in neuroscience.This study demonstrates that whole-brain functional connectivity patterns can be used to decode subject-driven cognitive states. By analyzing 90 functionally defined regions of interest (ROIs) across 14 large-scale resting-state brain networks, researchers developed a classifier that accurately identifies four cognitive states: undirected rest, episodic memory retrieval, serial subtraction, and silent singing. The classifier achieved 84% accuracy in a leave-one-out cross-validation and 85% accuracy in an independent cohort. It maintained high accuracy even with short imaging runs as brief as 30–60 seconds. Functional ROIs outperformed structural ROIs in classifying cognitive states across all tested durations. The study also tested the classifier's ability to reject a novel cognitive state, spatial navigation, and found it to be highly specific. The results suggest that whole-brain functional connectivity analysis can provide insights into naturalistic, continuous cognitive processing, offering a promising approach for decoding a wide range of subject-driven cognitive states from brief imaging data. The study highlights the potential of functional connectivity in understanding cognitive states without relying on explicit temporal markers, which is a significant advancement in neuroscience.