*4.2. Limitations*

There are some limitations to our study that should be noted. First, validation data are limited to the exercises of interest to our group. These vestibular exercises were chosen based on their common use and prescription to persons with ongoing balance problems after mTBI. Nonetheless, validation data in this study may not be generalizable to more dynamic tasks involving multiple planes of motion. Any estimation of trunk and head ROM and peak rotational velocities in more complex movements may, therefore, require further validation. Second, as accuracy of IMU output can be subject to sensor placement [13], it is possible that a different placement or a combined placement of sensors (i.e., an IMU on both the sternum and the lumbar spine), could provide a more accurate estimation of trunk motion. Third, our capture volume for optical motion data was limited, resulting in occasional marker occlusions when participants turned around the ends of our capture volume. Gap filling of missing optical markers and the resulting underestimation of head angles may be responsible for the fluctuations in error coincident with the turns. Fourth, errors in the trunk ROM and peak velocity may have been minimized had the trunk been defined differently. For U/D head motions, in particular, it is plausible that movement artifact occurred with neck movement pulling on the skin at the top of the sternum where one marker was placed. Finally, the IMU algorithm required the participants to stand in a static position so that the orientation of the sensors could be defined. This assumes that subjects can maintain a neutral stationary position, and further investigation into the effect of this potential issue is warranted.
