Many devices that offer lifestyle support, especially movement support, for the elderly have been recently developed. We have developed a new personal mobility device, Tread-Walk (TW), which is a walking support system for the elderly. Here we proposed a new control of a treadmill that enables users to walk with the same feeling as in normal level-ground walking. In this paper, we evaluated our proposed control method using a brain activity measurement, namely near-infrared spectroscopy. In particular, we quantitatively evaluated the operability, which has only been qualitatively evaluated previously. Results showed a notable tendency that brain activity depends on the accuracy of our proposed treadmill control. Then a possibility that a walking state could be evaluated from a brain activity was suggested. Our study validates the active velocity control of a treadmill according to brain activity measurement.