A noninvasive methodology, combining functional electrical stimulation and time-domain near-infrared spectroscopy (TD-NIRS), is developed to verify whether stroke-altered muscular metabolism on postacute patients. Seven healthy subjects and nine postacute stroke patients undergo a protocol of knee flex-extension induced by quadricep electrical stimulation. During the protocol, TD-NIRS measurements are performed on both rectus femoris to investigate whether significant differences arise between able-bodied and stroke subjects and between patients' paretic and healthy legs. During baseline, metabolic parameters do not show any significant differences among subjects. During stimulation, paretic limbs produce a knee angle significantly lower than healthy legs. During recovery, patients' healthy limbs show a metabolic behavior correlated to able-bodied subjects. Instead, the correlation between the metabolic behavior of the paretic and able-bodied legs allows the definition of two patients' subgroups: one highly correlated (R>0.87) and the other uncorrelated (R<0.08). This grouping reflects the patient functional condition. The results obtained on the most impaired patients suggest that stroke does not produce any systemic consequences at the muscle, but the metabolic dysfunction seems to be local and unilateral. It is crucial to enlarge the sample size of the two subgroups before making these preliminary results a general finding.