Various near-infrared spectroscopy (NIRS) variables have been used to estimate muscle lactate threshold (LT), but no study has determined which common NIRS variable best reflects muscle estimated LT. Establishing the inflection point of 2 regression lines for deoxyhaemoglobin (DeltaHHb(i.p).), oxyhaemoglobin (DeltaO2Hb(i.p.)), and tissue oxygenation index (TOIi.p.), as well as for blood lactate concentration, we then investigated the relationships between NIRS variables and ventilatory threshold (VT), LT, or maximal tissue hemoglobin index (nTHImax) during incremental cycling exercise. DeltaHHb(i.p.) and TOI(i.p.) could be determined for all 15 subjects, but DeltaO2Hb(i.p.) was determined for only 11 subjects. The mean absolute values for the 2 measurable slopes of the 2 continuous linear regression lines exhibited increased changes in 3 NIRS variables. The workload and VO2 at DeltaO2Hb(i.p.) and nTHImax were greater than those at VT, LT, DeltaHHb(i.p.), and TOI(i.p.). For workload and VO2, DeltaHHb(i.p.) was correlated with VT and LT, whereas DeltaO2Hb(i.p.) was correlated with nTHImax, and TOI(i.p.) with VT and nTHImax. These findings indicate that DeltaO2Hb strongly corresponds with local perfusion, and TOI corresponds with both local perfusion and deoxygenation, but that DeltaHHb can exactly determine deoxygenation changes and reflect O2 metabolic dynamics. The finding of strongest correlations between DeltaHHb and VT or LT indicates that DeltaHHb is the best variable for muscle LT estimation.