How Do You Prevent Decision Fatigue When Scanning Multiple Timeframe Setups Across Pairs?

Analyzing multi-timeframe charts (e.g., 1h bias down to 3m execution) across multiple instruments can quickly exhaust mental resources during active sessions. What specific systematic checklists or alert thresholds do you use to filter out low-probability noise and preserve cognitive focus?