The occurrence and severity of traumatic brain injury (TBI) in automotive crashes has remained a major issue. Since a mechanism of TBI has been understood as head rotational kinematics, the Brain Injury Criterion (BrIC) was developed, which assesses rotational velocity about each of three axes. The aim of this study is to investigate characteristics of head rotational kinematics and their effects on TBI metrics in full frontal and frontal oblique crashes. Head rotational kinematics of the THOR dummy were analyzed utilizing 120 cases, which consisted of 0° sled tests, 15° sled tests, full frontal rigid barrier vehicle crashes (FRB), and frontal offset obliquely-oriented moving deformable barrier vehicle crashes (OBL). Six degree-of-freedom head kinematics were applied to the average male model of the Global Human Body Model Consortium. Through finite element simulations, three tissue-level metrics for TBI were calculated, namely, cumulative strain damage measure (CSDM) for diffuse brain injury, maximum principal strain (MPS) for hemorrhage and contusion, and maximum bridging vein strain (MBVS) for acute subdural hematoma. Head flexion-extension motion (ω y ) was dominant in the 0°/FRB cases, while head twist motion (ω_z) was dominant in the 15°/OBL cases. BrIC values in the 15°/OBL cases were significantly higher than the 0°/FRB cases. CSDM and MPS showed fair correlations with ω_z (R² = 0.45 and 0.55, respectively), while MBVS was best correlated with ω_y (R² = 0.50). BrIC had a good correlation with CSDM and MPS (R² = 0.60 and 0.64, respectively), while its correlation with MBVS was weak (R² = 0.16). Compared to TBI risks based on BrIC values, the risks based on CSDM values were higher in the 0°/FRB cases and lower in the 15°/OBL cases. Additional analysis demonstrated that adjustment of the relative weighting of the head rotational velocity about each axis in the BrIC formula could improve the correlation of BrIC to TBI metrics. The results indicate that CSDM and MPS are affected by head rotation regardless of the axis, while MBVS is most affected by the flexion-extension motion. MBVS correlated significantly better with y-rotation in both crash categories, though the modeling of the bridging veins may affect this trend. BrIC was proved to be a fair predictor for MPS and CSDM in the studied datasets. However, it was shown that BrIC may not be robust to a wide range of TBI as well as a wide range of load cases. This study recommends further detail analysis on how various crash modes can have different sensitivities on TBI outcomes to establish a brain injury criteria.