Systems that can track and intelligently adapt to changes in users' cognitive capacities may help to improve human performance. The experiment reported here was designed to assess the benefit of placing neuro-physiological sensors on users to provide data about their cognitive states that can drive performance mitigations. A test group of 16 participants performed a primary task while monitoring for alerts (secondary task) under four conditions: no-mitigation (alerts presented as they arrive), random mitigation (system randomly alters the presentation times of alerts), sensor-driven mitigation (system uses sensor data to influence alert presentation times), and user-driven mitigation (participants pressed a button to influence the alert presentation times). Participants completed no-mitigation and user-driven mitigation conditions faster than sensor-driven and random mitigation conditions, but their accuracy scores did not differ significantly across the four conditions. This is likely due to a ceiling effect: participants' accuracy scores exceeded 90% in every condition. Future work should investigate the possibility that user-driven mitigation may, in some cases, improve performance better than sensor-driven mitigation.