Thoracic spine flexibility affects head motion, which is critical to control in motor vehicle crashes given the frequency and severity of head injuries. The objective of this study is to investigate the dynamic response of the human upper thoracic region. An original experimental/analytical approach, Isolated Segment Manipulation (ISM), is introduced to quantify the intact upper thoracic spine-pectoral girdle (UTS-PG) dynamic response of six adult post-mortem human subjects (PMHS). A continuous series of small displacement, frontal perturbations were applied to the human UTS-PG using fifteen combinations of speed and constraint per PMHS. The non-parametric response of the T1-T6 lumped mass segment was obtained using a system identification technique. A parametric mass-damper-spring model was used to fit the non-parametric system response. Mechanical parameters of the upper thoracic spine were determined from the experimental model and analyzed in each speed/constraint configuration. The natural frequencies of the UTS-PG were 22.9 ± 7.1 rad/sec (shear, n=58), 32.1 ± 7.4 rad/sec (axial, n=58), and 27.8 ± 7.7 rad/sec (rotation, n=65). The damping ratios were 0.25 ± 0.20 (shear), 0.42 ± 0.24 (axial), and 0.58± 0.32 (rotation). N-way analysis of variance (Type III constrained sum of squares, no interaction effects) revealed that the relative effects of test speed, pectoral girdle constraint, and PMHS anthropometry on the UTS-PG dynamic properties varied per property and direction. While more work is needed to verify accuracy in realistic crash scenarios, the UTS-PG model system dynamic properties could eventually aid in developing integrated anthropomorphic test device (ATD) thoracic spine and shoulder components to provide improved head kinematics and belt interaction.