Near-infrared spectroscopy (NIRS) signals quantify the oxygenated ($Δ$HbMbO 2 ) and deoxygenated ($Δ$HHbMb) heme group concentrations. $Δ$HHbMb has been preferred to $Δ$HbMbO 2 in evaluating skeletal muscle oxygen extraction because it is assumed to be insensitive to blood volume (BV) changes, but uncertainties exist on this assumption. To analyze this assumption a computational model of oxygen transport and metabolism is used to quantify the effect of O 2 delivery and BV changes on the NIRS signals from a canine model of muscle oxidative metabolism (Med.Sci.SportsExerc.,48(10)2013-2020,2016). The computational analysis accounts for microvascular ($Δ$HbO 2 , $Δ$HHb) and extravascular ($Δ$HMb, $Δ$HMb) oxygenated and deoxygenated forms. Simulations predicted muscle oxygen uptake and NIRS signal changes well for blood flows ranging from resting to contracting muscle. Additional NIRS signal simulations were obtained in the absence or presence of BV changes corresponding to a heme groups concentration changes ($Δ$HbMb=0-48$μ$M). Under normal delivery (Q=1.0L kg -1 min -1 ) of contracting muscle, capillary oxygen saturation (SO 2 ) was 62% with capillary $Δ$HbO 2 and $Δ$HHb of ±41$μ$M for $Δ$HbMb=0. An increase of BV ($Δ$HbMb =24mM) caused a $Δ$HbO 2 decrease (16mM) almost twice as much as the increase observed for $Δ$HHb (9$μ$M). When SO 2 increased to more than 80%, only $Δ$HbO 2 was significantly affected by BV changes. The analysis indicates that microvascular SO 2 is a key factor in determining the sensitivity of $Δ$HbMbO 2 and $Δ$HHbMb to BV changes. Contrary to a common assumption, the $Δ$HHbMb is affected by BV changes in normal contracting muscle and even more in the presence of impaired O 2 delivery.